leolca's blog

Segments Inventories and Syllable Inventories

2011-03-31T10:45:00.000-07:00

Another hypothesis is that the size of the segment inventory is related to the phonotactics of the language in such a way as to limit the total number of possible syllables that can be constructed from the segments and suprasegmental properties that it has. Languages might then have approximately equal numbers of syllables even though they differ substantially in the number of segments. Rough maintenance of syllable inventory size is evisaged as the functional of cyclic historical processes by, for example, Matisoff (1973). He outlines an imaginary language in which, at some arbitrary stating point, "the number of possible syllables is very large since there is a rich system of syllable-initial and -final consonants". At a later stage of the language these initial and final consonantal systems are found to have simplified but "the number of vowels has increased and lexically contrastive tones have arisen" maintaining contrasting syllabic possibilities. If tone or vowel contrasts are lost, consonat clustering will increase at the syllable margins again.

A brief investigation of the relationship between segmental inventory size and syllable inventory size was carried out by calculating the number of possible syllables in 9 languages. The languages are Tsou (418), Quechua (819), Thai (400), Rotokas (625), Gã (117), Hawaiian (424), Vietnamese (303), Cantonese, Higi, and Yoruba (the last three are not in UPSID but detailed data on the phonotactics are available in convenient form for these languages). The 9 languages range from those with small segment inventories (Rotokas, Hawaiian) to those with relatively large inventories (Vietnamese, Higi, Quechua) and from those with relatively simple suprasegmental properties (Tsou, Hawaiian, Quechua) to those with complex suprasegmental phenomena (Yoruba, Thai, Cantonese, Vietnamese). In calculating the number of possible syllables, general co-occirrence restrictions were taken into account, but the failure of a particular combination of elements to be attested if parallel combinations were permited is taken only as evidence of an accidental gap, and such a combination is counted as a possible syllable. The calculations reveal very different numbers of possible syllables in these languages. The totals are given in Table 1.5.

Table 1.5 Syllable inventory size of 9 selected languages
Language	Total possible syllables
Hawaiian	162
Rotokas	350
Yoruba	582
Tsou	968
Gã	2,331
Cantonese	3,456
Quechua	4,068
Vietnamese	14,430
Thai	23,638

Even with the uncertainties involved in this kind of counting, the numbers differ markedly enough for the conclusion to be drawn that language are not strikingly similar in terms of the size of their syllable inventories.

In following up this study, several tests were done to see which of a number of possible predictors correlated best with syllable inventory size. The predictors used were the number of segments, the number of vowels, the number of consonants, the number of permitted syllable structures (CV, CVC, CCVC, etc.), the number of suprasegmental contrasts (e.g. number of stress levels time number of tone), and a number representing a maximal count of segmental differences in which the number of vowels was multiplied by the number of suprasegmentals. Of these, the best predictor is the number of permitted syllable types (r = .69), an indication that the phonotactic possibilities of the language are the most important factor contributing to the number of syllables. The next best predictor is the number of suprasegmentas (r = .59), with the correlation with the various segmental counts all being somewhat lower. Although all the predictors tested show a positive simple correlation with the number of syllables, in a multiple regression analysis only the number of vowels contributes a worthwhile improvement to the analysis (r^2 change = .19) beyond the number of syllable types. Thus we can say that syllable inventory size does not depend heavily on segment inventory size. Nonetheless, because the predictors do have positive correlations with syllable inventory size, the picture is once
again of a tendency for complexity of different types to go together.

(Patterns of Sounds, Ian Maddieson)

Segments and Suprasegmentals

2011-03-31T10:41:00.000-07:00

Despite the failure to find any confirmation of a compensation hypothesis in several tests involving segmental subinventories, it is possible that the compensation exists at another level. One possibility was evidently in the minds of Firchow and Firchow (1969). In their paper on Rotokas (625), which has an inventory of only 11 segments, they remark that "as the Rotokas segmental phonemes are simple, the suprasegmentals are complicated". A similar view of a compensatory relationship between segmental and suprasegmental complexity seems implicit in much of the literature on the historical development of tone. For example, Hombert, Ohala and Ewan (1979) refer to "the development of contrastive tones on vowels because of the loss of a voicing distinction on obstruents". If this phenomenon is part of a pervasive relationship of compensation we would expect that, in general, languages with larger segmental inventories would tend to have more complex suprasegmental characteristics.

In order to test this predictions, the languages in UPSID which have less than 20 or more than 45 segments were examined to determine if the first group had obviously more complex patterns of stress and tone thatn the second. Both groups contain 28 languages. The findings on the suprasegmental properties of these languages, as far as they cam be ascertained, are summarized in Table 1.4.

Despite some considerable uncertainty of interpretation and the incompleteness of the data, the indications are quite clear that these suprasegmental properties are not more elaborate in the languages with simpler segmental inventories. If anything, they tend to be more elaborate in the languages with larger inventories.

There are more "large" languages with contrastive stress and with complex tone systems (more than 2 tones) that "small" languages. There are more "small" languages lacking stress and tone. The overall tendency appears once againn to be more that complexity of different kinds goes hand in hand, rather than for complexity of one sort to be balanced by simplicity elsewhere.

(Patterns of Sounds, Ian Maddieson)

Relationship between Size and Structure

2011-03-30T14:26:00.000-07:00

"The data in UPSID have been used to address the question of the relationship between the size of an inventory and its membership. The total number of consonants in an inventory varies between 6 and 95 with a mean of 22.8. The total number of vowels varies between 3 and 46 with a mean of 8.7. The balance between consonants and vowels within an inventory was calculated by dividing the number of vowels by the number of consonants. The resulting ratio varies between 0.065 and 1.308 with a mean of 0.402. The median value of this vowel ratio is about 0.36; in other words, the typical language has less than half as many vowels as it has consonants. There are two important trends to observe; larger inventories tend to be more consonant-dominated, but there is also a tendency for the absolute number of vowels to be larger in the languages with larger inventories. The first is shown by the fact that the vowel ratio is inversely correlated with the number of consonants in an inventory (r=-0.4, p=0.0001) and the second by the fact that the total of vowels is positively correlated with the consonant total (r=0.38, p=0.0001). However, a large consonant inventory with a small vowel inventory is certainly possible, as, for example, in Haida (700: 46C, 3V), Jaqaru (820: 38C, 3V) or Burushaski (915: 38C, 5V). Small consonant inventories with a large number of vowels seem the least likely to occur (cf. the findings of Hockett 1955), although there is something of an areal/genetic tendency in this direction in New Guinea languages such as Pawaian (612: 10C, 12V), Daribi (616: 13C, 10V) and Fasu (617: 11C, 10V). In these cases a small number of consonants is combined with a contrast of vowel nasality. Despite some aberrant cases, however, there is a general though weak association between overall inventory size and consonant/vowel balance: larger inventories tend to have a greater proportion of consonants."
(Patterns of Sounds, Ian Maddieson)

I made a graphic to show the relation between the number of vowels and consonants in a speech inventory. Each language in the UPSID is represented as a cross in the plot and the gray shading is the density of languages in the vowel-consonant plan.

Speech Inventories

2011-03-30T14:20:00.000-07:00

"Such an association suggests that inventory size and structure may be related in other ways as well. A simple form of such a hypothesis would propose that segment inventories are structured so that the smallest inventories contain the most frequent segments, and as the size of the inventory increases, segments are added in descending order of their overall frequency of occurrence. If this were so, all segments could be arranged in a single hierarchy. Such an extreme formulation is not correct, since no single segment is found in all languages. But if we add a corollary, that larger inventories tend to exclude some of the most common segments, then there is an interesting set of predictions to investigate. We may formulate these more cautiously in the following way: a smaller inventory has a greater probability of including a given common segment than a larger one, and a larger inventory has a greater probability of including an unusual segment type than a smaller one."
(Patterns of Sounds, Ian Maddieson)

I would say that there is a convergence towards a inventory size with a number of segments between 20 and 40 and using a restricted set o segments that tends to be common among languages those languages. As the language gets further away from this zone, we cannot say much about what would be its inventory.

Universality of Language

2011-03-29T13:28:00.000-07:00

Edwar Sapir wrote in 1921: "There is no more striking general fact about language than its universality. One may argue as to whether a particular tribe engages in activities that are worthy of the name of religion or of art, but we know of no people that is not possessed of a fully developed language. The lowliest South African Bushman speaks in the forms of a rich symbolic system that is in essence perfectly comparable to the speech of the cultivated Frenchman."

Shepard Tone

2011-03-16T06:24:00.000-07:00

Roger Shepard created an amusing auditory paradox, which is after him named: Shepard tone. It is based on a self-similar sequence of notes that consist on a superposition of 12 tones, each an octave higher then the lower neighbor. The 12 tones extend from the lower frequency limit of auditory perception to the upper frequency limit of hearing. In the present approach it goes from 10 Hz to 20,480 Hz, with all octave tones interleaved. The set of frequencies composing our Shepard tone is 10, 20, 40, 80, 160, 320, 640, 1280, 2560, 5120, 10,240, and 20,480 Hz. In order to create a paradoxal sound of a continouos ingreasing frequency sound, we create a sweep, an exponential chirp for each composing tone, going from its initial frequency to its following octave.

The instantaneous frequency of each tone is given by
$f(t) = f_0 k^t$.
As we want the final frequency after $T$ seconds to be $2 f_0$,
$f(T) = 2 f_0 = f_0 k^T$
That leads to
$k = e^{\frac{\ln 2}{T}}$
and the instantaneous frequency may be written as
$f(t) = f_0 e^{\frac{\ln 2}{T} t}$
An exponential chirp tone $x(t)$ is given by
$x(t) = \sin \left( 2 \pi \int_{0}^{t} f(\tau) d\tau \right)$
$x(t) = \sin \left( 2 \pi f_0 \frac{e^{\frac{\ln 2}{T}t}-1}{\ln 2 / T} \right)$

To acchieve a better result I have weighted the tones using a gaussian window, and I created the chirp tone with the double sampling frequency and in the end I made a resample to the desired sampling frequency, to attenuate the aliasing caused by the higher chirping tones.


fs = 44100;
fs2 = 2*fs;
ts = 1/fs2;
F = [10 20 40 80 160 320 640 1280 2560 5120 10240 20480];
w = gausswin(length(F));
x = [];
T = 10;
t = [0:1/fs2:T-1/fs2]';
x = zeros(size(t));
w(end+1)=0;
for i = 1 : length(F),
    x += linspace(w(i),w(i+1),length(t))' .* sin(2*pi*F(i)*(exp(log(2)/T * t)-1)/(log(2)/T));
end;
x = resample(x,1,2);

I used the function bellow to concatenate several chirping tones:


function y = concatenate(varargin)
y = varargin{1};
for i = 2 : nargin,
    x = varargin{i};
    o1 = linspace(1,0,11025)';
    o2 = linspace(0,1,11025)';
    y = [y(1:end-11025); o1.*y(end-11025+1:end)+o2.*x(1:11025); x(11025+1:end)];
end;

And the result is plotted in a Spectogram (see figure bellow) and might also be heard.


y = concatenate(x,x,x);
specgram(y,[],fs,[],[]);
soundsc(y,fs);

Power Law Exponent

2011-02-01T05:31:00.000-08:00

One main property of these fractals (or another way to express their main property, scalability) is that the ratio of two exceedances is going to be the ratio of the two numbers to the negative power of the power exponent.

Let us illustrate this. Say that you "think" that only 96 books a year will sell more than 250,000 copies (which is what happened last year), and that you "think" that the exponent is around 1.5. You can extrapolate to estimate that around 34 books will sell more than 500,000 copies -- simply 96 times (500,000/250,000)^(-1.5). We can continue, and note that around 8 books should sell more than a million copies, here 96 times (l,000,000/250,000)^(-1.5).

(...)

Table 2 illustrates the impact of the highly improbable. It shows the contributions of the top 1 percent and 20 percent to the total. The lower the exponent, the higher those contributions. But look how sensitive the process is: between 1.1 and 1.3 you go from 66 percent of the total to 34 percent. Just a 0.2 difference in the exponent changes the result dramatically -- and such a difference can come from a simple measurement error. This difference is not trivial: just consider that we have no precise idea what the exponent is because we cannot measure it directly. All we do is estimate from past data or rely on theories that allow for the building of some model that would give us some idea -- but these models may have hidden weaknesses that prevent us from blindly applying them to reality.

(The Black Swan, Nassim Nicholas Taleb)

Scale Invariance

2011-02-01T05:29:00.001-08:00

Emily Dickinson

2011-02-01T04:56:00.000-08:00

A word is dead

A word is dead
When it is said,
Some say.

I say it just
Begins to live
That day.

I'm Nobody! Who are you?

I'm Nobody! Who are you?
Are you -- Nobody -- Too?
Then there's a pair of us!
Don't tell! they'd advertise -- you know!

How dreary -- to be -- Somebody!
How public -- like a Frog --
To tell one's name -- the livelong June --
To an admiring Bog!

There is another sky

There is another sky,
Ever serene and fair,
And there is another sunshine,
Though it be darkness there;
Never mind faded forests, Austin,
Never mind silent fields -
Here is a little forest,
Whose leaf is ever green;
Here is a brighter garden,
Where not a frost has been;
In its unfading flowers
I hear the bright bee hum:
Prithee, my brother,
Into my garden come!

The Geometry of Nature

2011-02-01T04:27:00.000-08:00

Triangles, squares, circles, and the other geometric concepts that made many of us yawn in the classroom may be beautiful and pure notions, but they seem more present in the minds of architects, design artists, modern art buildings, and schoolteachers than in nature itself. That's fine, except that most of us aren't aware of this. Mountains are not triangles or pyramids; trees are not circles; straight lines are almost never seen anywhere. Mother Nature did not attend high school geometry courses or read the books of Euclid of Alexandria. Her geometry is jagged, but with a logic of its own and one that is easy to understand.

(The Black Swan, Nassim Nicholas Taleb)

The Great Asymmetry

2011-01-28T05:13:00.000-08:00

(...)

Indeed, the notion of asymmetric outcomes as the central idea of this book: I will never get to know the unknown since, by definition, it is unknown. However, I can always guess how it might affect me, and I should base my decisions around that.

This idea is often erroneously called Pascal's wager, after the philosopher and (thinking) mathematician Blaise Pascal. He presented it something like this: I do not know whether God exists, but I know that I have nothing to gain from being an atheist if he does not exist, whereas I have plenty to lose if he does. Hence, this justifies my belief in God.

Pascal's argument is severely flawed theologically: one has to be naïve enough to believe that God would not penalize us for false belief. Unless, of course, one is taking the quite restrictive view of a naive God. (Bertrand Russell was reported to have claimed that God would need to have created fools for Pascal's argument to work.)

But the idea behind Pascal's wager has fundamental applications outside of theology. It stands the entire notion of knowledge on its head. It eliminates the need for us to understand the probabilities of a rare event (there are fundamental limits to our knowledge of these); rather, we can focus on the payoff and benefits of an event if it takes place. The probabilities of very rare events are not computable; the effect of an event on us is considerably easier to ascertain (the rarer the event, the fuzzier the odds). We can have a clear idea of the consequences of an event, even if we do not know how likely it is to occur. I don't know the odds of an earthquake, but I can imagine how San Francisco might be affected by one. This idea that in order to make a decision you need to focus on the consequences (which you can know) rather than the probability (which you can't know) is the central idea of uncertainty. Much of my life is based on it.

(...)

(The Black Swan, Nassim Nicholas Taleb)

Backwards Narrative

2011-01-28T04:28:00.000-08:00

Philosophers since Aristotle have taught us that we are deep-thinking animals, and that we can learn by reasoning. It took a while to discover that we do effectively think, but that we more readily narrate backward in order to give ourselves the illusion of understanding, and give a cover to our past actions. The minute we forgot about this point, the "Enlightenment" came to drill it into our heads for a second time.

(The Black Swan, Nassim Nicholas Taleb)

We Just Can't Predict

2011-01-25T08:23:00.001-08:00

When I ask people to name three recently implemented technologies that most impact our world today, they usually propose the computer, the Internet, and the laser. All three were unplanned, unpredicted, and unappreciated upon their discovery, and remained unappreciated well after their initial use. They were consequential. They were Black Swans. O f course, we have this retrospective illusion of their partaking in some master plan. You can create your own lists with similar results, whether you use political events, wars, or intellectual epidemics.

You would expect our record of prediction to be horrible: the world is far, far more complicated than we think, which is not a problem, except when most of us don't know it. We tend to "tunnel" while looking into the future, making it business as usual, Black Swan-free, when in fact there is nothing usual about the future. It is not a Platonic category!

We have seen how good we are at narrating backward, at inventing stories that convince us that we understand the past. For many people, knowledge has the remarkable power of producing confidence instead of measurable aptitude. Another problem: the focus on the (inconsequential) regular, the Platonification that makes the forecasting "inside the box."

I find it scandalous that in spite of the empirical record we continue to project into the future as if we were good at it, using tools and methods that exclude rare events. Prediction is firmly institutionalized in our world. We are suckers for those who help us navigate uncertainty, whether the fortune-teller or the "well-published" (dull) academics or civil servants using phony mathematics.

(The Black Swan, Nassim Nicholas Taleb)

How Not to Bo a Nerd

2011-01-25T08:15:00.000-08:00

Think of a bookworm picking up a new language. He will learn, say, Serbo-Croatian or !Kung by reading a grammar book cover to cover, and memorizing the rules. He will have the impression that some higher grammatical authority set the linguistic regulations so that nonlearned ordinary people could subsequently speak the language. In reality, languages grow organically; grammar is something people without anything more exciting to do in their lives codify into a book. While the scholastic-minded will memorize declensions, the a-Platonic nonnerd will acquire, say, Serbo-Croatian by picking up potential girlfriends in bars on the outskirts of Sarajevo, or talking to cabdrivers, then fitting (if needed) grammatical rules to the knowledge he already possesses.

Consider again the central planner. As with language, there is no grammatical authority codifying social and economic events; but try to convince a bureaucrat or social scientist that the world might not want to follow his "scientific" equations. In fact, thinkers of the Austrian school, to which Hayek belonged, used the designations tacit or implicit precisely for that part of knowledge that cannot be written down, but that we should avoid repressing. They made the distinction we saw earlier between "know-how" and "know-what"—the latter being more elusive and more prone to nerdification.

To clarify, Platonic is top-down, formulaic, closed-minded, self-serving, and commoditized; a-Platonic is bottom-up, open-minded, skeptical, and empirical.

(The Black Swan, Nassim Nicholas Taleb)

Phony Philanthropy

2011-01-19T05:37:00.000-08:00

Frédéric Bastiat was a nineteenth-century humanist of a strange variety, one of those rare independent thinkers—independent to the point of being unknown in his own country, France, since his ideas ran counter to French political orthodoxy (he joins another of my favorite thinkers, Pierre Bayle, in being unknown at home and in his own language). But he has a large number of followers in America.

In his essay "What We See and What We Don't See," Bastiat offered the following idea: we can see what governments do, and therefore sing their praises—but we do not see the alternative. But there is an alternative; it is less obvious and remains unseen.

Recall the confirmation fallacy: governments are great at telling you what they did, but not what they did not do. In fact, they engage in what could be labeled as phony "philanthropy," the activity of helping people in a visible and sensational way without taking into account the unseen cemetery of invisible consequences. Bastiat inspired libertarians by attacking the usual arguments that showed the benefits of governments. But his ideas can be generalized to apply to both the Right and the Left.
Bastiat goes a bit deeper. If both the positive and the negative consequences of an action fell on its author, our learning would be fast. But often an action's positive consequences benefit only its author, since they are visible, while the negative consequences, being invisible, apply to others, with a net cost to society. Consider job-protection measures: you notice those whose jobs are made safe and ascribe social benefits to such protections. You do not notice the effect on those who cannot find a job as a result, since the measure will reduce job openings. In some cases, as with the cancer patients who may be punished by Katrina, the positive consequences of an action will immediately benefit the politicians and phony humanitarians, while the negative ones take a long time to appear — they may never become noticeable. One can even blame the press for directing charitable contributions toward those who may need them the least.

(The Black Swan, Nassim Nicholas Taleb)

Patterns vs. Randomness

2011-01-18T03:55:00.000-08:00

There is another, even deeper reason for our inclination to narrate, and it is not psychological. It has to do with the effect of order on information storage and retrieval in any system, and it's worth explaining here because of what I consider the central problems of probability and information theory.

The first problem is that information is costly to obtain.

The second problem is that information is also costly to store—like real estate in New York. The more orderly, less random, patterned, and narratized a series of words or symbols, the easier it is to store that series in one's mind or jot it down in a book so your grandchildren can read it someday.

Finally, information is costly to manipulate and retrieve.

(...)

Consider a collection of words glued together to constitute a 500-page book. If the words are purely random, picked up from the dictionary in a totally unpredictable way, you will not be able to summarize, transfer, or reduce the dimensions of that book without losing something significant from it. You need 100,000 words to carry the exact message of a random 100,000 words with you on your next trip to Siberia. Now consider the opposite: a book filled with the repetition of the following sentence: "The chairman of [insert here your company name] is a lucky fellow who happened to be in the right place at the right time and claims credit for the company's success, without making a single allowance for luck," running ten times per page for 500 pages. The entire book can be accurately compressed, as I have just done, into 34 words (out of 100,000) ; you could reproduce it with total fidelity out of such a kernel. By finding the pattern, the logic of the series, you no longer need to memorize it all. You just store the pattern. And, as we can see here, a pattern is obviously more compact than raw information. You looked into the book and found a rule. It is along these lines that the great probabilist Andrey Nikolayevich Kolmogorov defined the degree of randomness; it is called "Kolmogorov complexity."

We, members of the human variety of primates, have a hunger for rules because we need to reduce the dimension of matters so they can get into our heads. Or, rather, sadly, so we can squeeze them into our heads. The more random information is, the greater the dimensionality, and thus the more difficult to summarize. The more you summarize, the more order you put in, the less randomness. Hence the same condition that makes us simplify pushes us to think that the world is less random than it actually is.

And the Black Swan is what we leave out of simplification.

(The Black Swan, Nassim Nicholas Taleb)

Perception is biologically bounded

2011-01-18T03:51:00.000-08:00

Actually, as I am writing this, there is news of a pending lawsuit by a patient going after his doctor for more than $200,000 — an amount he allegedly lost while gambling. The patient claims that the treatment of his Parkinson's disease caused him to go on wild betting sprees in casinos. It turns out that one of the side effects of L-dopa is that a small but significant minority of patients become compulsive gamblers. Since such gambling is associated with their seeing what they believe to be clear patterns in random numbers, this illustrates the relation between knowledge and randomness. It also shows that some aspects of what we call "knowledge" (and what I call narrative) are an ailment.

Once again, I warn the reader that I am not focusing on dopamine as the reason for our overinterpreting; rather, my point is that there is a physical and neural correlate to such operation and that our minds are largely victims of our physical embodiment. Our minds are like inmates, captive to our biology, unless we manage a cunning escape. It is the lack of our control of such inferences that I am stressing. Tomorrow, someone may discover another chemical or organic basis for our perception of patterns, or counter what I said about the left-brain interpreter by showing the role of a more complex structure; but it would not negate the idea that perception of causation has a biological foundation.

(The Black Swan, Nassim Nicholas Taleb)

Diversification vs. Unification

2010-12-13T09:55:00.000-08:00

Fortunately, if a condition of vocabulary balance does exist in a given sample of speech, we shall have little difficulty in detecting it because of the very nature and direction of the two Forces involved. On the one hand, the Force of Unification will act in the direction of decreasing the number of different words to 1, while increasing the frequency of that 1 word to 100%. Conversely, the Force of Diversification will act in the opposite direction of increasing the number of different words, while decreasing their average frequency of occurrence towards 1. Therefore number and frequency will be the parameters of vocabulary balance.

Since the number of different words in a sample of speech together with their respective frequencies of occurrences can be determined empirically, it is clear that our next step is to seek relevant empiric information about the number and frequency of occurrences of words in some actual samples of speech.

(George K. Zipf, Human Behavior and the Principle of Least Effort, 1949)

Double superlatives

2010-12-13T07:08:00.001-08:00

Yet when we offer a prize to the submarine commander who sinks the greatest number of ships in the shortest possible time, we have a double superlative -- a maximum number and a minimum time -- which renders the problem completely meaningless and indeterminate, as becomes apparent upon reflection. Double superlatives of this sort, which are by no means uncommon in present-day statements, can lead to a mental confusion with disastrous results.

As pointed out years ago, the frequent statement, "in a democracy we believe in the greatest good for the greatest number," contains a double superlative and therefore is meaningless and indeterminate. (In Part Two we shall see that the distribution of goods and services are in fact governed by a single superlative.) Intimately connected with the "singleness of the superlative" is what might be called the singleness of the objective whose implications are often overlooked (i.e., the pursuit of one objective may preclude or frustrate the pursuit of the scoond objective). These two concepts apply to all studies in ecology.

(George K. Zipf, Human Behavior and the Principle of Least Effort, 1949)

Mandelbrot Set

2010-12-06T08:39:00.000-08:00

The Mandelbrot set M is defined by a family of complex quadratic polynomials
$P_c:\mathbb C\to\mathbb C$
given by
$P_c: z\mapsto z^2 + c$, where c is a complex parameter.

If the sequence $(0, P_c(0), P_c(P_c(0)), P_c(P_c(P_c(0))), \ldots)$ (starting with $z=0$) does not escapes to infinity, the complex number c is said to belong to the set.

A simple Octave/Matlab code was created to plot the complex numbers on the Mandelbrot set.


mset=[];
x=[-2:0.001:1];
y=[-1:0.001:1];
k=0;
for k1=1:length(x),
for k2=1:length(y),
    c=x(k1)+i*y(k2);
    z=0;
    for it=1:100,
        z = z^2+c;
    end;
    if(z < Inf), k++; mset(k)=c; end;
end;
end;
figure; plot(mset,'k.');

Simulation on Sand Avalanches

2010-11-30T08:06:00.000-08:00

It is common to observe how complex systems in nature present similar patterns, what is due to the self-organized criticality. Critical points work as attractors towards which systems evolve seeking for equilibrium. A common example are the sand piles. It is observed that as more grain of sand are dropped, new avalanches may occur. The avalanches may be of different proportions, local small avalanches or big ones, spreading widely in the pile. The observation of these avalanches shows that they follow a Zipf's law.

Here I present the results of a simulation of sand avalanches. The video shows a set of samples, which with 10 continuous seconds taken every 80 seconds from the simulation. At every second a new sand is randomly dropped on the board. If a certain peak has 4 more grain of sand then any of its neighbours, an avalanche happens towards this neighbour, what might lead to a cascade of avalanches.

The histogram bellow show the frequency of occurrence of avalanches of different sizes, described as: 0 (no avalanche), 1 (avalanche happened just on the level the grain was dropped), 2 (avalanche spreads to the next level), and so forth.

If we present the result above as a log-log plot of the types of avalanches (according to their magnitude) ranked according to their frequency of occurrence, we get to the figure bellows, which clearly is a power law, a Zipf's law. The type of avalanche with rank one is the most frequent, the type ranked two is the second most frequent, and so on. The plot bellow shows the rank vs. frequency of occurrence. The numbers presented near the curve (1 2 3 0 4 5 6 ...) are the type of avalanche: 1 stands for avalanche only in the level where the grain is dropped; 2 when the avalanche spreads to the next level; 3 when it spread two levels bellow; 0 when there is no avalanche; etc.

Zipf's Law

2010-11-26T12:43:00.000-08:00

Zipf's statistical law is based on the idea that two "opposing forces" are in constant operation in a system. In the stream of speech, they are: the Force of Unification and the Force of Diversification. Any speech is is a result of the interplay of these forces, that through a self-organizing process reaches a critical point, a point of "balance" between them. This balance is observed on the relation between the frequency of occurrence of words (f) and their rank (k), their product is constant.

$f(k;s,N)=\frac{1/k^s}{\sum_{n=1}^N (1/n^s)}$.

In the formula above, s stands for the exponent that characterizes the distribution; and N stands for the number of elements in the set. The formula states that the frequency of occurrence of a given element is given by the rank of this element within a given, which distribution is characterized by s. The figure bellow presents how the relationship of frequency and rank is when plotted in a log-log scale for different values of s.

Zipf developed the idea using an intrinsic linguistic or psychological reason to explain this phenomena observed in the world of words. He named his theory the "Principle of Least Effort" to explain why frequently encountered words are chosen to be shorter in order to require a little mental and physical effort to recall them and utter/write them. According to Alexander et al. (1998), Zipf’s law seems to hold regardless the language observed. "Investigations with English, Latin, Greek, Dakota, Plains Cree, Nootka (an Eskimo language), speech of children at various ages, and some schizophrenic speech have all been seen to follow this law"(Alexander et al., 1998).

The Zipf’s law is also observed in other phenomena, for example: the magnitude of earthquakes (it is common to have many small earthquakes, but big ones are rare) (Abe and Suzuki, 2005); the population in cities (there are few megalopolis, but thousands of small cities) (Gabaix, 1999); the distribution of total liabilities1 of bankrupted firms in high debt range(Fujiwara, 2004); the number of requests for webpages(Adamic and Huberman, 2002); etc.

The observation of languages also points to a Zipf’s law, what makes one more evidence that languages operate in a rather random way. Performing a statistical analysis of language means to acknowledge its unpredictable nature, without which there would be no communication at all. The analysis of languages as a statistical process has advantages over the qualitative analysis, it "is able to afford to neglect the narrow limits of one language and concentrate on linguistic problems of a general character" (Trnka, 1950). Although this conflict between randomness and rationality might rise suspicious on the character of languages, Miller wisely pointed: "If a statistical test cannot distinguish rational from random behavior, clearly it cannot be used to prove that the behavior is rational. But, conversely, neither can it be used to prove that the behavior is random. The argument marches neither forward nor backward" (Miller, 1965).

References:

Abe, S. and Suzuki, N. (2005). Scale-free statistics of time interval between successive earthquakes. Physica A: Statistical Mechanics and its Applications, 350(2-4):588–596.

Adamic, L. A. and Huberman, B. A. (2002). Zipf’s law and the internet. Glottometrics,
3:143–150.

Alexander, L., Johnson, R., and Weiss, J. (1998). Exploring zipf’s law. Teaching Mathe
matics Applications, 17(4):155–158.

Fujiwara, Y. (2004). Zipf law in firms bankruptcy. Physica A: Statistical and Theoretical Physics, 337(1-2):219–230.

Gabaix, X. (1999). Zipf’s law for cities: An explanation. Quarterly Journal of Economics, 114(3):739–67.

Miller, G. A. and Taylor, W. G. (1948). The perception of repeated bursts of noise. The Journal of the Acoustical Society of America.

Trnka, Bohumil (1950). Review of: G.K.Zipf, The psychobiology of language. Human behavior and the principle of least effort.

The minimum wage miracle

2010-11-20T06:49:00.000-08:00

With the new Constitution of 1988, it was granted to every worker, the right to earn at least an amount called "minimum wage". In the second chapter, sixth article, it defines this minimum wage as the amount capable of attending the basic vital necessities of a worker and his family with home, food, education, health, leisure, clothing, hygiene, transportation and social security. It is also establish, that this minimum wage should undergoes periodic readjustments to overcome inflationary degradation of the worker's purchasing power. Although the Constitution does not specifies how often those readjustments should happen and what amount should be given, we observe in the history of readjustments that there is a regularity. There is almost one readjustment every year and the readjusting factor kept following the rule: CPI (consumer price index, which is used as a measure of inflation) + percent GDP growth. The rule of readjustment followed by the government in the last years to define the readjustment factor is: the CPI of the year that has just passed plus the percent GDP growth experienced in the year before. That means: readjustment_factor[n] = CPI[n-1]*GDP[n-2]. This might be observed in the figure right bellow, where the actual minimum wage value is plotted in blue and the predicted readjustment (using the rule) in red:

It is clear now that the readjustment try to follow this rule.

The interesting point about this readjustment rule is that it promotes the distribution of wealth. As the country keeps growing, the minimum wage also keeps increasing, at a larger step. Not everyone wage keeps increasing at the minimum wage's rate. It might clearly be apprised in the figure bellow. The cyan curve shows the average wage of Brazilians. If it were to be readjusted using the same rule the minimum wage uses, it would gives us the green curve. Looking at the graphic, we see that the average wage is always bellow the green curve, what shows that, on average, the wages don't follow the minimum wage increase.

The consequence of this is better viewed when we normalized all curves by the actual minimum wage value, what is presented in the next figure.

In 1995, the average Brazilian wage was around 8 times greater than the minimum wage. In the last year, in 2009, it reached a ratio 3:1. This means distribution of wealth. Observing the curve, it seems like the curve is going to saturate at a ratio between 3:1 and 2:1, what would leave Brazil in a similar position to developed countries. Here is some ratios of average wage to minimum wage in some countries: USA 3.28:1, Spain 2.55:1, Netherlands 2.22:1, Portugal 2.02:1, France 2.00:1, UK 1.88:1 (data from: source 1 and source 2).

How to deceive dumb people with exponential growth

2010-11-19T10:39:00.000-08:00

It is pretty easy to deceive an outsider that the last few years have made a tremendous better profit in comparison to the previous ones.

When we experience a exponential growth we get surprised to see how the growth was tremendous in the last few samples. In fact, on average, the percent growth is the same, but what is important (mainly in political matters) is to convince another that we are experiencing glorious times.

A simple analysis of various examples show us percent growing rates almost constant. That is natural. Observe the GDP growth of a developing country, or the growth on the exportation, or the growth of a large company. Usually you will find a percent growing rate that is almost constant.

If we consider than a growing rate to be a random variable with a certain mean and variance, we might sketch the growing curve. Bellow I present some simulations where the percent growing rate was modeled as a normally distributed random variable with mean 4 and some distinct variances: 4, 6, 8, and 10. The growing curves are presented bellow:

And here another sketch, where the graphics are separated according to their variances.

Lets take one graphic as an example. It is plotted with its percent growth shown bellow, along with a moving average of the percent growth.

If you juts look at the first plot, you would conclude that after time 60 and/or 80 we got a bigger improvement. But you may look things in a different way. If you look at the second subplot, you may observe the percent growth through time. Now it is easy to see that it is almost the same through time, and the first picture is clearly deceiving.

Another way to see how deceiving our first graphic was is to plot the data in a logarithm scale, as shown right bellow:

Don't believe straight away on your eyes, or you might get deceived. Don't believe on exploding graphics jumping out of the paper, they are made to deceive you. Don't believe on millions and billions on a moth of a politician. To see the truth, you have to get your hands dirty.

Self-Organized Criticality and Gaia

2010-11-18T18:13:00.000-08:00

In a seminal work, Jim Lovelock, an English scientist, came up with a fascinating idea that all life on earth can be viewed as a single organism. This idea has stuck many scientists as preposterous since it flies in the face of the usual reductionist approach and smacks of New-Age philosophy. Lovelock's idea is that the environment, including the air that we breathe, should not be viewed as an external effect independent of biology, but that it is endogenous to biology. The oxygen represents one way for species to interact. Lovelock
noted that the composition of oxygen has increased dramatically since life originated. The oxygen content is far out of equilibrium. The layer of ozone, an oxygen molecule, that protects life on earth did not just happen to be there, but was formed by the oxygen created by life itself. Therefore, it does not make sense to view the environment, exemplified by the amount of oxygen, as separate from biological life. One should think of the earth as one single system.

What does it mean to say that the earth is one living organism? One might ask in general: What does it mean that anything, such as a human, is one organism? An organism maybe defined as a collection of cells or other entities that are coupled to each other so that they may exist, and cease to exist, at the same time -- that is, they share one another's fate. The definition of what represents an organism depends on the time scale that we set. ln a time scale of 100 million years, all humans represent one organism. At short time scales, an ant's nest is an organism. There is no fundamental difference between genetically identical ants carrying material back and forth to build and operate their nest, and genetically identical human cells organizing themselves in structures and sending blood around in the system to build and operate a human body.

Thus a single organism is a structure in which the various parts are interconnected, or "functionally integrated" so that the failure of one part may cause the rest of the structure to die, too. The sand pile is an organism because sand grains toppling anywhere may cause toppling of grains anywhere in the pile.

One might think of self-organized critically as the general, underlying theory Ear the Gaia hypothesis In the critical state the collection of species represents a single coherent organism following its own evolutionary dynamics. A single triggering event can cause an arbitrarily large fraction of the ecological network to collapse, and eventually be replaced by a new stable ecological network. This would be a "mutated" global organism. At the critical point all species influence each other. ln this state they act collectively as a single meta-organism, many sharing the same fate. This is highlighted by the very existence of large-scale extinctions A meteorite might have directly impacted a small part of the organism, but a large Fraction of the organism eventually died as a result.

Within the SOC picture. the entire ecology has evolved into the critical state. It makes no sense to view the evolution of individual species independently, Atmospheric oxygen might be thought of as the bloodstream connecting the various parts of our Gaia organism, but one can envision organisms that interact in different ways.

The vigorous opposition to the Gaia hypothesis, which represents a genuine holistic view of life, represents the frustration of a science seeking to maintain its reductionist view of biological evolution.

(from: How Nature Works: The Science of Self-Organized Criticality, Per Bak)

Here followa two interesting videos with Lovelock.

evolving towards dumbness

2010-11-17T11:22:00.000-08:00

Sewall Wright, an American geneticist known for his influential work on evolutionary theory, proposed the idea of fitness landscapes. The fitness describes a certain genotype or phenotype performance in the environment. As a species evolves, its fitness changes. As the environment changes, the specie's fitness also changes. When you reach a peak in the fitness landscapes, there is no way to achieve higher peak, but go through a valley. In the last century we clearly reached a peak in the development of human intelligence. Now we experience a dumbing down process. Dumbness is spread everywhere. Dumbness procreates faster then intelligence. As result of this evolving process, the world is going dumb. We are going down the hill.

Financial assets and Zipf law

2010-11-12T11:08:00.000-08:00

According to Mandelbrot, the variation on commodities and financial assets through time follow a Zipf law. He made a analysis with data from cotton price over a few months. I decided to try it myself and see if it really holds.

I got some financial assets data and observed the frequency of occurrence of variation within certain ranges. The graphics bellow present this result. In the first graphic, the variation (%) range was linearly sliced. In the second graphic, I made a logarithmic spaced slices. The third image has a linearly spaced slices, but the number of slices is quite smaller.

When the number of slices is smaller, we clearly see a straight line, what means a power law holds. When the number of slices is bigger, it behaves like a line with a saturation for very frequent values.

Petrobras

Bovespa index

Banco do Brasil

"But economics is like sand, not like water. Decisions are discrete, like the grains of sand, not continuous, like the level of water. There is friction in real economics, just like in sand. We don't bother to advertise and take our apples to the market when the expected payoff of exchanging a few apples and oranges is too small. We sell and buy stocks only when some threshold price is reached, and remain passive in between, just as the crust of the earth is stable until the force on some asperity exceeds a threshold. We don't continually adjust our holdings to fluctuations in the market. In computer trading, this threshold dynamics has been explicitly programmed into our decision pattern.
Our decisions are sticky This friction prevents equilibrium from being reached, just like the friction of sand prevents the pile from collapsing to the flat state. This totally changes the nature and magnitude of fluctuations in economics.

Economists close their eyes and throw up their hands when it comes to discussing market fluctuations, since there cannot be any large fluctuations in equilibrium theory: "Explanations for why the stock market goes up or down belong on the funny pages," says Claudia Goldin, an economist at Harvard. If this is so, one might wonder, what do economists explain?

The various economic agents follow their own, seemingly random, idiosyncratic behavior. Despite this randomness, simple statistic patterns do exist in the behavior of markers and prices. Already in the 1960s, a few years before his observations of fractal patterns in nature, Benoit Mandelbrot analyzed data for fluctuations of the prices of cotton and steel stocks and other commodities. Mandelbrot plotted a histogram of the monthly variation of cotton prices. He counted how many months the variation would be 0.1% (or
-0.1% ), how many months it would be 1%, how many months it would be 10%, etc. He found a "Levy distribution" for the price fluctuations. The important feature of the Levy distribution is that it has a power law tail for large events, just like the Gutenberg-Richter law for earthquakes. His findings have been largely ignored by economists, probably because they don't have the faintest clue as to what is going on.

Traditionally, economists would disregard the large fluctuations, treating them as "atypical" and thus not belonging in a general theory of economics. Each event received its own historical account and was then removed from the data set. One crash would be assigned to the introduction of program trading, another to excessive borrowing of money to buy stock. Also, they would "detrend" or "cull" the data, removing some long-term increase or decrease in the market. Eventually they would end up with a sample showing only small
fluctuations, but also totally devoid of interest. The large fluctuations were surgically removed from the sample, which amounts to throwing the baby outwith the bathwater. However, the fact that the Iarge events follow the same behavior as the small events indicates that one common mechanism works for all scales -- just as for earthquakes and biology.

How should a generic model of an economy look? Maybe very much like the punctuated equilibrium model for biological evolution described in Chapter 8. A number of agents (consumers, producers, governments, thieves, and economists, among others) interact with each other. Each agent has a limited ser of options available. He exploits his options in an attempt to increase his happiness (or "utility function" as the economists call it to sound more scientific), just as biological species improve their fitness by mutating. This affects other agents in the economy who now adjust their behavior to the new situation. The weakest agents in the economy are weeded out and replaced with other agents. or they modify their strategy, for instance by copying more successful agents.

This general picture has not been developed yet. However, we have constructed a simplified toy model that offers a glimpse of how a truly interactive, holistic theory of economics might work."
(How Nature Works: The Science of Self-Organized Criticality, Per Bak)

Pseudoscience

2010-11-08T05:56:00.000-08:00

"As pointed out by the philosopher Karl Popper, prediction is out best means of distinguishing science from pseudoscience. To predict the statistics of actual phenomena rather than the specific outcome is a quite legitimate and ordinary way of confronting theory with observarions". (How Nature Works: The Science of Self-Organized Criticality, Per Bak)

search tips

2010-10-14T10:51:00.000-07:00

Basic
- "+" - Result must contain word
- "-" - Result must not contain word
- "OR" and "|" - Applied between two words, it will find "this or that", or both. The "OR" operator must be uppercase and have a space between the 2 words on each side. The "|" operator does not need a space between the words
- " "" " - Finds an exact match of the word or phrase
- "~" - Looks for synonyms or similar items. Eg: "~run" will match runner's and marathon
- ".." - Indicates that there's a range between number. Eg: 100..200 or $100..$200
- "*" - Matches a word or more. Eg: "Advanced * Form" finds "Advanced Search Form"
- "word-word" - All forms (spelled, singe word, phrase and hyphenated

Important
- "site:" - Search only one website or domain. Eg: "PC site:wazem.org" will find PC within wazem.org
- "filetype:" or "ext:" - Search for docs in the file type. Eg: "Linux tutorial filetype:pdf" will find Linux tutorial in the pdf format
- "link:" - Find linked pages (pages that point to the URL)
- "define:" - Provides definition for a word or a phrase
- "cache:" - Display Google's cached version of a web page.
- "info:" - Info about a page
- "related:" - Websites related to the URL
- "allinurl:" - All words must be in the URL
- "allintitle:" - All words must be in the title of the page
- "intittle:" - Match words in the title of the page
- "source:" - News articles from a specific source

Calculations
- "+ - * /" - Normal math signs. Eg: 12 * 4 + 2 - 1 /2
- "% of" - Percentage. Eg:10% of 100
- "^" or "**" - Raise to a power
- units "in" units - Convert Units (currency, measurements, weight). Eg: 300 lbs in Kg, 40 in hex

Others
- "book" or "books" - Search books. Eg: book "LPI Linux Certification in a Nutshell"

spoken english statistics

2010-08-24T12:13:00.000-07:00

In order to approximate the frequency of occurrence of phones in spoken english, I used the Gutenberg Project database to get written texts and the CMU Pronouncing Dictionary to get a phonetic transcription of those words.

I used the top 100 books on the list of Gutenberg database. With them I cound build a list of 179,044 types and 14,144,013 tokens. Just to state a comparison, "the Second Edition of the 20-volume Oxford English Dictionary contains full entries for 171,476 words in current use, and 47,156 obsolete words. To this may be added around 9,500 derivative words included as subentries"(see the reference). That builds a total of 228,132 entries. So have around 78% of the entries of the Oxford Dictionary.

With this data at hand and some perl scripting, I made the following lists and ordered them by the frequency of occurrence:

1. list of words;

2. list of phones;

3. list of diphones;

4. list of triphones;

5. list of quadriphones;

6. an interface to browse the data (click on the image bellow);

7. log-log graphic of word rank vs. word frequency;

8. log-log graphic of phone rank vs. phone frequency;

Exponential fit of the data:

Semi-logy plot:

9. log-log graphic of diphone rank vs. diphone frequency;

10. log-log graphic of triphone rank vs. triphone frequency;

11. log-log graphic of quadriphone rank vs. quadriphone frequency;

12. frequency of occurrence of a phone given that a certain phone occurred before;

13. frequency of occurrence of a phone given that a certain phone occurs after;

14. Conditiona probability of occurrence of a given phone given that another has occurred.

14. the kullback-leibler distance (relative entropy) between the phones in english and a uniform distributed random variable is: 0.48363 bits;

15. considering as a dissimilarity measure between two phones the sum of their individual frequency of occurrence minus the frequency of occurrence of the diphone with this pair of phones, we get the following dissimilarity matrix and we peform a MDS of the data, as shown bellow.

zoom:

16. Words letters-length

17. Frequency of occurrence of words with a certain letters-length normalized by the number of possible permutations of letters with repetition with the same length.

18. Average number of letters in a word across word's rank

19. Words phones-length

20. Frequency of occurrence of words with a certain phonemic-length normalized by the number of possible permutations of phones with repetition with the same length.

21. Average number of phones in a word across word's rank

22. Cumulative probability of phones. The 8 first most frequent phones ([ə, t, n, s, ɪ, r, d, l]) account for half of all phones occurrences in the data.

23. Here are present two types of graphics to verify the contribuition of words frequency to the final phones frequency. The upper plot shows the occurrence of words with a certain phone. The lower one show an extimation of the probability of occurrence of a certain phone across the rank of words.

semantic web

2010-08-22T09:59:00.001-07:00

I have just made my first prototype of a semantic web. :)

First I listed the occurences of words placed just by a certain word. The list bellow shows the occurence of words adjacent to the portuguese word 'casa'.


era : 41
dono : 35
minha : 35
sua : 30
nossa : 30
estava : 30
verde : 28
foi : 25
porta : 24
dona : 21
velha : 19
me : 18
esta : 18
rua : 18
dele : 18
entrou : 18
ir : 17
dela : 17
chegou : 16
noite : 15
fora : 13
ia : 12
...

Using this list and the list of the words in this list I built a semantic web!

See other examples:
ainda
depois
ele
quando
tudo
casa
disse
era
tempo

word frequency

2010-08-17T10:08:00.000-07:00

Here is the process I made to create a frequency list of portuguese words using all the text from Machado de Assis available at http://machado.mec.gov.br/. In the total, this database has 1.645.474 tokens and 62.809 types.

First download all pdfs.


mkdir pdf
cd pdf
wget http://machado.mec.gov.br/arquivos/pdf/romance/marm01.pdf
wget http://machado.mec.gov.br/arquivos/pdf/romance/marm02.pdf
wget http://machado.mec.gov.br/arquivos/pdf/romance/marm03.pdf
...

Then convert everything into text.


for file in $( ls pdf/*.pdf ); 
do echo $file; outfile=${file//pdf/txt}; 
pdftotext -enc UTF-8 $file $outfile; 
done

Then you just need to run my perl script to get the list of words and their occurancy.


#!/usr/bin/perl
my $dirname = $ARGV[0];
my %count_of;
opendir(DIR, $dirname) or die "can't opendir $dirname: $!";
while (defined($filename = readdir(DIR))) {
   open (FILE, $dirname . $filename);
   while () {
      chomp;
      $_ = lc $_;
      $_ =~ s/\d+/ /g; # remove all numbers
      $_ =~ s/[^a-zA-Z0-9_áéíóúàãõâêôçü]+/ /g;
      #$_ =~ s/\xC3//g; # remove strange one
      foreach my $word ( split /\s+/, $_){
          $count_of{$word}++;
      }
   }
   close (FILE);
}
closedir(DIR);
print "All words and their counts: \n";
foreach $value (sort {$count_of{$b} <=> $count_of{$a} } keys %count_of)
{
  print "$value : $count_of{$value}\n";
}

Get the script here.

Voilà! And here is the result!


All words and their counts: 
a : 75485
que : 69366
de : 66929
o : 61165
e : 57056
não : 34354
se : 28067
do : 25059
um : 24125
da : 21992
os : 19764
é : 18307
uma : 16521
em : 15381
com : 14954
as : 14793
para : 13114
mas : 12390
lhe : 11922
me : 10966
ao : 10962
era : 10340
por : 10266
no : 10114
mais : 9148
na : 9003
à : 8719
como : 8506
dos : 7669
eu : 6972
ou : 6696
ele : 6310
foi : 5445
das : 5305
há : 5215
nem : 5169
sem : 4387
quando : 4283
disse : 4140
já : 3924
ela : 3815
ser : 3774
nos : 3687
tudo : 3537
ainda : 3514
só : 3402
depois : 3358
tempo : 3137
casa : 3098
...

Get the complete list here.

Combination

2010-07-18T13:50:00.000-07:00

Here is a simple function for Octave/MatLab I wrote to create the combinations of the numbers in a vector. Suppose you want to get the possible combinations of the numbers [1 2 3 4] arrenged in 3, what would give you [1 2 3], [1 2 4], [1 3 4] and [2 3 4]. You just have to use the function bellow calling X = combinations([1 2 3 4],3), and X will be the matrix with all combinations.


function X = combinations(x,k)
%
%  X = combinations(x,k)
%  Create all combinations (without repetition) of the
%  elements in x arragend in groups of k items.
%  example:
%  x=[1 2 3 4];
%  X = combinations(x,3)
%  X =
%     1   2   3
%     1   2   4
%     1   3   4
%     2   3   4
%

if(size(x,1) > size(x,2)), x = x'; end;
n = length(x);
X = [];
if(k == 1),
  X = x';
else
for l = 1 : n-k+1,
    C = nchoosek(n-l,k-1);
    xtemp = x;
    xtemp(1:l) = [];
    X = [X;  [repmat(x(l),C,1) combinations(xtemp,k-1)] ];
end;
end;

Holiday

2010-07-15T05:42:00.000-07:00

(taken from www.phdcomics.com)

Subtitles on PS3

2010-07-14T08:16:00.000-07:00

Unfortunately the only way I found, until now, to play downloaded videos with subtitles on my PS3 is using a tool called AVIAddXSubs. This tool create a DivX video by adding a source video (.avi, .mpg. etc) and its subtitle (.srt). The subtitle is added to the DivX container (the DivX Media Format (DMF) has support to multiple subtitles, multiple audio tracks and multiple video streams, among other things, just like Matroska). Although AVIAddXSubs is a Windows program, it might run on Linux, thanks to Wine. I have just tried it and it did work! I could get my video playing on my PS3 with subtitles. :)

models

2010-07-01T07:21:00.000-07:00

(...) The value of a model is that often it suggests a simple summary of the data in terms of the major systematic effects together with a summary of the nature and magnitude of the unexplained of random variation. (...)

Thus the problem if looking intelligently at data demands the formulation of patterns that are thought capable of describing succinctly not only the systematic variation in the data under study, but also for describing patterns in similar data that might be collected by another investigator at another time and in another place.

(...) Thus the very simple model
\[ y = \alpha x + \beta ,\]
connecting two quantities y and x via the parameter pair (α,β), defines a straight-line relationship between y and x. (...) Clearly, if we know α and β we can reconstruct the values of y exactly from those if x (...). In practice, of course, we never measure the ys exactly, so that the relationship between y and x is only approximately linear. (...)

The fitting of a simple linear relationship between the ys and the xs requires us to choose from the set of all possible pairs of parameters values a particular pair (a, b) that makes the patterned set $\hat{y}_1,\ldots,\hat{y}_n$ closest to the observed data. In order to make this statement precise we need a measure of 'closeness' or, alternatively, of distance or discrepancy between the observed ys and the fitted $\hat{y}$s. Examples of such discrepancy functions include the $L_1$-norm

\[S_1(y,\hat{y}) = \sum | y_i - \hat{y}_i | \]
and the $L_\infty$-norm
\[S_\infty(y,\hat{y}) = \max_i | y_i - \hat{y}_i | .\]

Classical least squares, however, chooses the more convenient $L_2$-norm or sum of squared deviations
\[S_2(y,\hat{y}) = \sum ( y_i - \hat{y}_i )^2 \]
as the measure of discrepancy. These discrepancy formulae have two implications. First, the straightforward summation of individual deviations, either $| y_i - \hat{y}_i |$ or $( y_i - \hat{y}_i )^2$, each depending on only one observation, implies that the observations are all made on the same physical scale and suggests that the observations are independent, or at least that they are in some sense exchangeable, so justifying an even-handed treatment of the components. Second, the use of arithmetic differences $y_i - \hat{y}_i$ implies that a given deviation carries the same weight irrespective of the value of $\hat{y}$. In statistical terminology, the appropriateness of $L_p$-norms as measures of discrepancy depends on stochastic independence and also on the assumption that the variance of each observation is independent of its mean value. Such assumptions, while common and often reasonable i practive, are by no means universally applicable.

(Generalized Linear Models, P. McCullagh and J.A. Nelder)

Veja desenvolve novo método para testar indenpência de duas variáveis aleatórias.

2010-06-19T07:43:00.001-07:00

Na edição de 16 de Junho de 2010, a Veja mostra um novo método para testar indenpência de duas variáveis aleatória, no caso em questão, aplicado para analisar a independência das Eleições e Copa do Mundo.

in reference to: Eleições: Equilíbrio inédito na corrida presidencial - Edição 2169 - Revista VEJA (view on Google Sidewiki)

Create your own booklets

2010-06-14T13:53:00.001-07:00

Here is a easy way to create your own booklets and print'em... easy and fast.
You just have to put the paper, run the script, and change the paper when asked to.
(*note: there is no need to reorder the pages... the script does it automatically)

Here is a script to do it all:


#!/bin/bash
file=$1
pdf2ps $file /tmp/${file%%.pdf}.ps
psbook /tmp/${file%%.pdf}.ps /tmp/out.ps
psnup -2 /tmp/out.ps /tmp/out2up.ps
ps2pdf /tmp/out2up.ps ${file%%.pdf}-booklet.pdf
echo Do you want to print the booklet now? \(y\)es or \(n\)no:
read PRINT
if [ "$PRINT" = "y" ]; then
   lp -o page-set=odd ${file%%.pdf}-booklet.pdf
   echo change paper now! \(no need to reorder pages\)
   read PRINT
   lp -o page-set=even -o outputorder=reverse ${file%%.pdf}-booklet.pdf
fi

you may download it here.

pdfposter

2010-06-02T13:17:00.001-07:00

Go to official pdfposter website here.

This tool is very simple and good. You can split a PDF document into several to create a poster. See the example bellow:


pdfposter -p 4x4a4 /tmp/drawing.pdf /tmp/out.pdf

input: 1 page, A0, pdf file:
drawing.pdf

output: 16 pages, A4, pdf file:
out.pdf

But there is a problem... printers usually have a printable area, so you need to add a margin to your document in order to don't lose anything when you print it. Here is how I did it for the example above.

First, use pdfposter to create a smaller version (160x247mm, that means 2.5mm of top, bottom, left and right margins).

pdfposter -m160x247mm -pA0 drawing.pdf out.pdf

After that, create a TeX document that will include this pdf, place each page inside a A4 page and include cropping marks. Here is the source code:

\documentclass{article}
% Support for PDF inclusion 
\usepackage[final]{pdfpages}
% Support for PDF scaling
\usepackage{graphicx}
\usepackage[dvips=false,pdftex=false,vtex=false]{geometry}
\geometry{
   paperwidth=160mm,
   paperheight=247mm,
   margin=2.5mm,
   top=2.5mm,
   bottom=2.5mm,
   left=2.5mm,
   right=2.5mm,
   nohead
}
\usepackage[cam,a4,center,dvips]{crop}
\begin{document}
% Globals: include all pages, don't auto scale
\includepdf[pages=-,pagecommand={\thispagestyle{plain}}]{/tmp/out.pdf}
\end{document}

And... voilà! Here is the final pdf. And it page will be like this:

A Formal Theory of Syntax

2010-04-27T13:27:00.000-07:00

(Zellig Harris, Language and Information)

1.1 Problems and Methods

To Consider the structure of language, especially its syntax -- that is, how sentences are built from words -- we note first the usual approaches. Scientists coming to the problem from the outside often seek regularities in the sequential relation among the words in a sentence, since the data presents itself sequentially. However, sufficient regularities have not been found, for reasons that will appear later. In contrast, people who work with language analyze it o n the basis of what is called grammatical relations, such as the subject and object of a verb, or the relation of an affix to its host word. Here too there are difficulties. Few if any grammatical relations appear in the same way in all languages, so the individual relations cannot be taken as primitives for language as such. The situation is rather that in each language there are some relations that can be called grammatical, but a satisfactory general definition is lacking. Furthermore, grammatical relations are unique to natural language, and if we can describe language only in such terms we will be unable to compare language with anything else in the world, not even with such close relatives as gesture on the one hand and mathematics on the other. Finally, the elements on which grammatical relations hold are not adequately defined. The one type of element that is precisely established is the set of phonemes, the characteristic sounds of a language ; indeed the discovery of phonemes is the beginning of a precise science of language. But as to words, if they are thought of as correlations of sound sequences with meanings, we are left with many problems, such as homonyms (as in see, sea, and the Holy See), or with the two pronunciations of economics, not to mention many exceptional situations. And as to sentences, the lack of a general definition is well known. Hence, while traditional grammars can provide adequate descriptions of a language, they do not supply a framework for considering the structure of language in general.

1.2 Procedures Yielding the Elements

To see how the description of language structure is achieved, we note first how the elements can be established.

First, it is possible to determine the phonemic distinctions in a language by a behavioral test that does not involve the specific meaning of words or the investigator's judgment of phonetic similarity. The test consists of one speaker of a language uttering, in random order, repetitions of two words (e.g., see and sea, or hard and heart) while the hearer (another speaker) judges which pronunciation are repetitions of which. (...) These discriminations create sound types, as against merely a scientist's aggregation of sound tokens, and the discriminations themselves constitute the discrete and definite ("phonemic") elements of of which everything else in language is constructed. (...)Phonemes are obtained by the most economical way of collecting into one element phonemic distinctions having different environments (...).

Next, it is possible to locate word boundaries within utterances of a language (and morpheme boundaries, e.g., affixes, within words) by a stochastic process -- that is, a process that checks the n+1th item given the first n items. (...) Each point at which the number of different possible next phonemes (or letters) peaks, i.e., at which the number is greater than immediately before or after, is (in most cases) a word or morpheme boundary. Such peaking arises because not all phoneme sequences make words and not all word sequences make sentences.

Finally, it is possible to locate the boundary of sentences within utterances. When sequences of words in utterances are studied, it is found that to a certain extent one can classify the local combinations into required ones and several kinds of possible ones. A complex stochastic process on the word sequence in an utterance, in respect to these types of successor classification, reveals a periodicity: at certain points, the successor possibilities return to the situation at the beginning of the utterance. These recurrent-event points segment the utterance into a succession of structurally independent sentences.

The stochastic process just mentioned are important even for a known language, where we know from experience what are the words and the sentences. First, the processes show that words and sentences exist, not merely by cultural convention or by some semantic properties but by restrictions of combinations, that is of occurrences relative to each other, in the physical components of speech. Secondly, they show that each type of entity is definable as a relation holding among entities of a smaller (more local) type. The phoneme sequence relation that makes words proves to be of little interest: it does not in general tie up with the other properties of words -- not with their meanings and not with the word sequence relation that makes sentences. However, the word sequence relations are explicit and are of decisive importance for the structure and meaning of sentences. It will be seen that these last relations are primarily a matter of the frequency of words relative to each other in utterances, or sentences, of a language.

1.3 Syntax Procedures

(...)

We are now ready to consider what combinations of words occur in the language, in contrast with those that do not. This cannot be done by simply listing them. First, the list would be too vast. Second, the set of sentences is not well defined: there are many marginal sentences about which speakers are not certain or do not agree about whether they are said at all, or are in the language. Third, languages changes, and no list would be correct over a sufficient period. Instead of listing, therefore, we try to find what constraints preclude the combinations that are not in the language, what restrictions affect the equiprobability of word occurrences in respect to each other in utterances of the language. It will be seen that there are three types of constraints on word combinations that make sentences, and that each one carries a type of meaning, so that the meaning of a sentence is determined directly from the words and the constraints. The three are: a partial ordering that creates sentencehood, a probability inequality that allows for word meaning, and a reducing of phonemic forms that does not affect the objective meaning.

1.4 The Partial-Order Constraint

The first constraint creates sentences structures. It is a partial order of words, that is (roughly) an ordering in which some words are higher or lower on some scale than others, while some are neither higher nor lower than others. The partial order holds between word occurrences in utterances. It determines all sentences, but it overt only in a subset from which, however, all other sentences can be derived. Grammatical relations can be defined in terms of it.

(...)

The partial order is a constraint on word combinations: it says that in the argument position next to a given operator the frequency (or probability) of certain words -- those not in the argument class for that operator -- is zero. Each satisfaction of the partial order, i.e., each word sequence in which all the source words have their requirement satisfied, is a sentence. (...) The partial-order relation has a meaning: as will be seen later, each operator is being said about its argument, so that the meaning of the partial order is roughly predication.

While the dependence is based on the word combinations in a particular body of data, it is intended to predict the combinations in any utterance of the language except insofar as transformations (the reductions in the third constraint) alter the shapes (and apparent presence) and positions of words.

Now, this dependence relation has an important property. If we ask what determines for each word which word class it requires as argument, we find that the required words are identified by what they in turn require. (...)

(...)

There must be, in the language and in each sentence, at least one zero-level argument that requires nothing, for otherwise one couldn't have any words in a sentence. There must also be at least one first-level operator that requeries only words that require null, for nothing else could enter a sentence after the zero-level words (...). And there would have to be second-level operators, at least one of whose requirements is a
first-level operator, if we are to have any sentences beyond the elementary ones.

Thus the relations that imposes the partial order is not just the dependence of word on a stated class of words, but the dependence of a word on the dependence property of words. This is the kind of relation that can define a system without recourse to any externally defined elements; it has the property of a mathematical object. Having come to this, it is worth noting that the language elements involved, namely words, have indeed no inherent property that has to be used for sentences construction. The sound of words are not related to their meaning or their combinability, and are even dispensed with in writing (especially in pictographic writing). Even the meanings of words, as will be noted in the third lecture, are in part determined from their combinations rather than purely from their identity. From the viewpoint of the partial order, then, the word occurrences from a set of arbitrary elements closed under the dependence-on-dependence relation, with every combination that satisfies this relation being a sentence.

The importance of this relations will be clearer when it is seen that almost all further class and operations in language, and almost all language meanings, are formulated on the constructions resulting from this relation. It should be noted that the operator-argument relation produced by this dependence has important similarities to functions in categorical grammar logic. The differences arise from the different purposes in a syntax of logic and a syntax of natural language.

1.5 The Likelihood Constraint

We have obtained the gross structure of sentences. It is still necessary to describe how a particular word is chosen for a sentence, how certain combinations are more likely than others. This is done by a second constraint, on word likelihood. Whereas the first constraint creates sentences structure, the second specifies word meanings. It does not necessarily create meaning, since many words with their meanings must have been in use singly before being used in sentences, but it specifies meaning to any detail desired, and it enables a word to extend its meaning, and to have different meanings in different operator-argument environment. The first constraint set probability = 0 for words outside the required class in argument position; this leaves room for any probability > 0 for words of the required class. Nothing says that all words must have equal frequency in respect to their operator or argument, or that the frequency must be random, or must fluctuate. In fact we find in language that each word has a particular and roughly stable likelihood of occurring as argument, or operator, with a given other word, though there are many cases of uncertainty, disagreement among speakers, and change through time. These roughly stable likelihoods, and especially the selection frequency, which will be mentioned in a moment, conform to and fix the meanings of words.

We speak here of likelihood under an operator (or over an argument), in the sense of estimated frequency or probability per fixed number of occurrences of that operator (or argument); no one has actually counted the frequencies of various words in argument position under another word. Nevertheless it should be noted that counting
such frequencies over a small sample of the language is not as impossibly vast a task as it might seem to be, and this because we are not speaking of frequency in respect to other words in arbitrary sentences but only in the word pairs or triples in operator-argument relation, which is the elementary sentential structure and the sentential component of all sentences, and which constitutes the great bulk of meaning-characterizing, roughly stable relative frequencies.

Each word has a somewhat fuzzy selection of other words that are more likely than average to occur in the position for its argument -- that is, more likely that would
be expected if the occurrences were random or equal in frequency. (...)

In addition there exist words with exceptionally high likelihood, and this on several different grounds. A word may have high likelihood as a total of many ordinary likelihoods, if it is in the selection of exceptionally many operators. (...)

Some high-likelihood word occurrences are recognized to be such by the very fact that they have been reduced. (...)

There are also words with exceptionally low likelihood in particular situations. (...)

1.6 The Reduction Constraint

The third constraint makes existing sentences more compact. It consists, for each language, of a few specifiable types of reduction, even to zero, in the phonemic shape of particular word occurrences. First, the domain of reduction: what is reducible is the high-likelihood (or otherwise favored) material. Certain words that have exceptionally high likelihood or special status in a given position are reducible. (...) The words that have highest likelihood, i.e., are expectable in a given environment contribute little or no information when they enter there, in the information-theoretic sense It is relevant that reduction takes place in several different high-likelihood and special status situations. This suggests that what determines reducibility is not simply high frequency but low information, which is the common property of all of these situations. Note that the ability of the hearer to supply the zeroed word shows that the to-be-zeroed occurrence of the word carried no further information that had to be given by the speaker.

This suggestion is supported by the fact that words that have exceptionally low likelihood in a particular operator environment ca, when they do occur there, block reductions that would otherwise take place there.

1.8 Properties of the Base

We have seen the constraints on word combinarion: the partial order of word dependence that created sentence structure, the likelihood inequalities that fit word meanings, the reduction of high-likelihood word occurrences, and finaly the linearizations. Each acts on the resultants of its predecessor. The constraints partition the set of sentences into two major sets. Without reduction, they create a base set from which all other sentences are derived. What is important here is that neither the base set not the other set, the derived (reduced) set, is merely a residue of the other. On one hand, the structure of the base set is not just a description of all those sentences that could not be derived from something. On the other hand, the derivations are not just any change needed to obtain the remaining senteces from the base set. Rather, the base set and the reductions each have simple and understandable structures on their own terms, and it is a notrivial result that the whole set of sentences is characterized by just these two structures.

Rational vs. Real

2010-04-26T07:14:00.001-07:00

Unicode decimal and hex numbers for IPA symbols

2010-04-07T13:32:00.000-07:00

Alphabetic

font: http://www.unicode.org/charts/
(excluding the standard characters a-z)

Symbol	decimal	hex	value
ɑ	593	0251	open back unrounded
ɐ	592	0250	open-mid schwa
ɒ	594	0252	open back rounded
æ	230	00E6	raised open front unrounded
ɓ	595	0253	vd bilabial implosive
ʙ	665	0299	vd bilabial trill
β	946	03B2	vd bilabial fricative
ɔ	596	0254	open-mid back rounded
ɕ	597	0255	vl alveolopalatal fricative
ç	231	00E7	vl palatal fricative
ɗ	599	0257	vd alveolar implosive
ɖ	598	0256	vd retroflex plosive
ð	240	00F0	vd dental fricative
ʤ	676	02A4	vd postalveolar affricate
ə	601	0259	schwa
ɘ	600	0258	close-mid schwa
ɚ	602	025A	rhotacized schwa
ɛ	603	025B	open-mid front unrounded
ɜ	604	025C	open-mid central
ɝ	605	025D	rhotacized open-mid central
ɞ	606	025E	open-mid central rounded
ɟ	607	025F	vd palatal plosive
ʄ	644	0284	vd palatal implosive
ɡ	609	0261	vd velar plosive (but the IPA has ruled that an ordinary g is also acceptable)
ɠ	608	0260	vd velar implosive
ɢ	610	0262	vd uvular plosive
ʛ	667	029B	vd uvular implosive
ɦ	614	0266	vd glottal fricative
ɧ	615	0267	vl multiple-place fricative
ħ	295	0127	vl pharyngeal fricative
ɥ	613	0265	labial-palatal approximant
ʜ	668	029C	vl epiglottal fricative
ɨ	616	0268	close central unrounded
ɪ	618	026A	lax close front unrounded
ʝ	669	029D	vd palatal fricative
ɭ	621	026D	vd retroflex lateral
ɬ	620	026C	vl alveolar lateral fricative
ɫ	619	026B	velarized vd alveolar lateral
ɮ	622	026E	vd alveolar lateral fricative
ʟ	671	029F	vd velar lateral
ɱ	625	0271	vd labiodental nasal
ɯ	623	026F	close back unrounded
ɰ	624	0270	velar approximant
ŋ	331	014B	vd velar nasal
ɳ	627	0273	vd retroflex nasal
ɲ	626	0272	vd palatal nasal
ɴ	628	0274	vd uvular nasal
ø	248	00F8	front close-mid rounded
ɵ	629	0275	rounded schwa
ɸ	632	0278	vl bilabial fricative
θ	952	03B8	vl dental fricative
œ	339	0153	front open-mid rounded
ɶ	630	0276	front open rounded
ʘ	664	0298	bilabial click
ɹ	633	0279	vd (post)alveolar approximant
ɺ	634	027A	vd alveolar lateral flap
ɾ	638	027E	vd alveolar tap
ɻ	635	027B	vd retroflex approximant
ʀ	640	0280	vd uvular trill
ʁ	641	0281	vd uvular fricative
ɽ	637	027D	vd retroflex flap
ʂ	642	0282	vl retroflex fricative
ʃ	643	0283	vl postalveolar fricative
ʈ	648	0288	vl retroflex plosive
ʧ	679	02A7	vl postalveolar affricate
ʉ	649	0289	close central rounded
ʋ	651	028B	vd labiodental approximant
ʊ	650	028A	lax close back rounded
ʌ	652	028C	open-mid back unrounded
ɣ	611	0263	vd velar fricative
ɤ	612	0264	close-mid back unrounded
ʍ	653	028D	vl labial-velar fricative
χ	967	03C7	vl uvular fricative
ʎ	654	028E	vd palatal lateral
ʏ	655	028F	lax close front rounded
ʑ	657	0291	vd alveolopalatal fricative
ʐ	656	0290	vd retroflex fricative
ʒ	658	0292	vd postalveolar fricative
ʔ	660	0294	glottal plosive
ʡ	673	02A1	vd epiglottal plosive
ʕ	661	0295	vd pharyngeal fricative
ʢ	674	02A2	vd epiglottal fricative
ǀ	448	01C0	dental click
ǁ	449	01C1	alveolar lateral click
ǂ	450	01C2	alveolar click
ǃ	451	01C3	retroflex click

The Price of a Publication

2010-03-24T06:22:00.000-07:00

After long months or years of research we want to publish our results and show the community what we have done. Lets take an example: suppose we made some research on Antennas and Propagation, so we definitely want to publish on the IEEE Transactions on Antennas and Propagation. If we want to get published me must pay for it ($110 per printed page) and we must give up all copyrights.

(source: http://ieeeaps.org/aps_trans/howto.htm)

Too bad... but there is no other way. :(

But the process doesn't stop there. We want to publish, we need to read other papers and we need people to read our paper. Is it free? Nooooo! Of course not. If you wanna read a paper on the IEEE Transactions on Antennas and Propagation, you gotta pay for it $30.

(source: http://ieeexplore.ieee.org/articleSale/Sarticle.jsp?arnumber=5075644)

Wow! And what IEEE makes with all that money???

IEEE is just an example... many (or maybe all) other Journals make the same.

Fonts in Ubuntu

2010-03-23T04:13:00.000-07:00

Some cool free fonts you may find here. After downloading and extracting the ttf files, create a folder inside '/usr/share/fonts/truetype' where you want to store your fonts, for example: 'sudo mkdir /usr/share/fonts/truetype/myfonts'. Then just copy the ttf files, 'sudo cp *.ttf /usr/share/fonts/truetype/myfonts/' and run the following 'sudo fc-cache -f -v'. Now it is ready to use!

The Method of Triadic Comparisons

2010-03-18T13:54:00.000-07:00

(Application of multidimensional scaling to subjective evaluation of coded speech, Joseph L. Hall)

A problem with metric multidimensional scaling in human listening experiments is that listeners are required to associate numbers with dissimilarities. Different listeners use numbers differently, and it is difficult for a listener to use numbers consistently throughout the course of a listening test. A test in which listeners are required to rank order dissimilarities, so that judgments are of the form ``greater than'' or ``less than,'' is more satisfactory in this respect. Tests of this sort can be analyzed by nonmetric multidimensional scaling, in which the numbers in the dissimilarity matrix are a rank ordering of the distance between objects. In our example, the two cities that are closest together receive a rank ordering of one, and the two cities that are furthest apart receive a rank ordering of 45. With 10 cities there are 45 intercity distances. The criterion used by the multidimensional scaling program is that the rank ordering of distances in the stimulus space agree with the rank ordering of input dissimilarities. According to Young and Harris op. cit., p.
127 , the nonmetric minimization problem is much more difficult than the metric problem and requires an iterative solution. In our airline mileage example, the result of applying nonmetric classical multidimensional scaling to this dissimilarity matrix is an object space that differs only slightly from the first one.

Rank ordering the intercity mileages is easy; it involves simply sorting 45 numbers. In practice, when a listener is asked to make judgment about the relative dissimilarities of auditory stimuli, rank ordering more than a few stimuli becomes very time consuming and puts unacceptable demands on the listener's memory. A technique that has been developed to solve this problem is the method of triadic comparisons. Rather than being presented with all $n$ stimuli and being asked to rank order their dissimilarities, the subject is presented with stimuli three at a time and is asked to judge which two of the three are most dissimilar and which two of the three are most similar. The most dissimilar pair is given a score of two, the most similar pair is given a score of zero, and the remaining pair is given a score of one. This process is repeated for all $\binom{n}{3}$ triads of the n stimuli taken three at a time, and the scores resulting from each triad are added up to obtain the dissimilarity matrix. An advantage of this method is that each trial is completely self-contained. The subject's judgment on a given trial is based only on the three stimuli that are presented on that trial. This differs from MOS testing, in which the subject's judgment on a given trial is influenced in an uncontrolled manner by stimuli presented on other trials. The method of triadic comparisons does not allow for direct comparison of all $\binom{n}{2}$ stimulus pairs, so that with sparsely populated stimulus spaces some distortion is possible. In our flying-mileages example, the stimulus space resulting from the method of triadic comparisons differs from the original stimulus space; but in spite of the drastically modified task, the basic structure remains unchanged.

(...)

The solution space generated by weighted multidimensional scaling is not rotatable: the dimensions of the stimulus space and the subject space are dictated by the data and are not arbitrary. This is an extremely important property. It means that we can ask listeners to make judgments about similarities or differences among speech samples without specifying what features they should base these judgments on; and, to the extent that the assumptions of the model are valid, we can determine the relevant features from the experimental results.

A prova de que os Jedis existem

2010-03-18T08:39:00.001-07:00

Só falta mostrar como o campo induzido por essas ondas alphas podem influenciar as ondas de outra pessoa nas proximidade. Provavelmente o gesto com a mão possui também algum efeito.

Quantitative Expression for lnformation

2010-03-17T09:52:00.000-07:00

(Transmission of Information, R. V. L. Hartley)

At each selection there are available three possible symbols. Two successive selections make possible $3^2$, or 9, different permutations or symbol sequences. Similarly n selections make possible $3^n$ different sequences. Suppose that instead of this system, in which three current values are used, one is provided in which any arbitrary number $s$ of different current values can be applied to the line and distinguished from each other at the receiving end. Then the number of symbols available at each selection is $s$ and the number of distinguishable sequences is $s^n$.

Consider the case of a printing telegraph system of the Baudot type, in which the operator selects letters or other characters each of which when transmitted consists of a sequence of symbols (usually five in number). We may think of the various current values as primary symbols and the various sequences of these which represent characters as secondary symbols. The selection may then be made at the sending end among either primary or secondary symbols. Let the operator select a sequence of $n_2$ characters each made up of a sequence of $n_1$, primary selections. At each selection he will have available as many different secondary symbols as there are different sequences that can result from making $n_1$ selections from among the $s$ primary symbols. If we call this number of secondary symbols $s_2$, then

\[s_2 = s^{n_1}\]

For the Baudot System

\[s_2 = 2^5 = 32 \textmf{characters}\]

The number of possible sequences of secondary symbols that can result from $n_2$ secondary selections is

\[s_2^{n_2} = s^{n_1 n_2}\]

Now $n_1 n_2$ is the number $n$ of selections of primary symbols that would have been necessary to produce the same sequence had there been no mechanism for grouping the primary symbols into secondary symbols. Thus we see that the total number of possible sequences is $s^n$ regardless of whether or not the primary symbols are grouped
for purposes of interpretation.

This number $s^n$ is then the number of possible sequences which we set out to find in the hope that it could be used as a measure of the information involved. Let us see how well it meets the requirements of such a measure.

For a particular system and mode of operation $s$ may be assumed to be fixed and the number of selections $n$ increases as the communication proceeds. Hence with this measure the amount of information transmitted would increase exponentially with the number of selections and the contribution of a single selection to the total information transmitted would progressively increase. Doubtless some such increase does often occur in communication as viewed from the psychological standpoint. For example, the single word "yes" or "no", when coming at the end of a protracted discussion, may have an extraordinarily great significance. However, such cases are the exception rather than the rule. The constant changing of the subject of discussion, and even of the individuals involved, has the effect in practice of confining the cumulative action of this exponential relation to comparatively short periods.

Moreover we are setting up a measure which is to be independent of psychological factors. When we consider a physical transmission system we find no such exponential increase in the facilities necessary for transmitting the results of successive selections. The various primary symbols involved are just as distinguishable at the receiving end for one primary selection at for another. A telegraph system finds one ten-word message no more difficult to transmit than the one which preceded it. A telephone system which transmits speech successfully now will continue to do so as long as the system remains unchanged. In order then for a measure of information to be of practical engineering value it should be of such a nature that the information is proportional to the number of selections. The number of possible sequences is therefore not suitable for use directly as a measure of information.

We may, however, use it as the basis for a derived measure which does meet the practical requirements. To do this we arbitrarily put the amount of information proportional to the number of selections and so choose the factor of proportionality as to make equal amounts of information correspond to equal numbers of possible sequences.
For a particular system let the amount of information associated with $n$ selections be

\[H = K n\]

where $K$ is a constant which depends on the number $s$ of symbols available at each selection. Take any two systems for which $s$ has the values $s_1$ and $s_2$ and let the corresponding constants be $K_1$ and $K_2$. We then define these constants by the condition that whenever the numbers of selections $n_1$ and $n_2$, for the two systems are such that the number of possible sequences is the same for both systems, than the
amount of information is also the same for both; that is to say, when

\[s_1^{n_1} = s_2^{n_2}\]

\[H = K_1 n_1 = K_2 n_2 \]

from which

\[ \frac{K_1}{log s_1} = \frac{K_2}{log s_2} \]

This relation will hold for all values of $s$ only if $K$ is connected with $s$
by the relation

\[K = K_0 log s\]

where $K_0$ is the same for all systems. Since $K_0$ is arbitrary, we may omit it if we make the logarithmic base arbitrary. The particular base selected fixes the size of the unit of information. Putting this value of $K$ in (4),

\[H = n log s\]
\[H = log s^n\]

What we have done then is to take as our practical measure of information the logarithm of the number of possible symbol sequences.

The situation is similar to that involved in measuring the transmission loss due to the insertion of a piece of apparatus in a telephone system. The effect of the insertion is to alter in a certain ratio the power delivered to the receiver. This ratio might be taken as a measure of the loss. It is found more convenient, however, to take the logarithm of the power ratio as a measure of the transmission loss.

If we put $n$ equal to unity, we see that the information associated with a single selection is the logarithm of the number of symbols available; for example, in the Baudot System referred to above, the number $s$ of primary symbols or current values is 2 and the information content of one selection is $log 2$; that of a character which involves 5 selections is $5 log 2$. The same result is obtained if we regard a character as a secondary symbol and take the logarithm of the number of these symbols, that is, $log 2^5$, or $5 log 2$. The information associated with 100 characters will be $500 log 2$. The numerical value of the information will depend upon the system of logarithms used. Increasing the number of current values from 2 to say 10, that is, in the ratio 5, would increase the information content of a given number of selections in the ratio $\frac{log 10}{log 2}$, or 3.3. Its effect on the rate of transmission will depend upon how the rate of making selections is affected. This will be discussed later.

When, as in the case just considered, the secondary symbols all involve the same number of primary selections, the relations are quite simple. When a telegraph system is used which employs a non-uniform code they are rather more complicated. A difficulty, more apparent than real, arises from the fact that a given number of secondary or character selections may necessitate widely different numbers of primary selections, depending on the particular characters chosen. This would seem to indicate that the values of information deduced from the primary and secondary symbols would be different. It may easily be shown, however, that this does not necessarily follow.

If the sender is at all times free to choose any secondary symbol, he may make all of his selections from among those containing the greatest number of primary symbols. The secondary symbols will then all be of equal length, and, just as for the uniform code, the number of primary symbols will be the product of the number of characters by the maximum number of primary selections per character. If the number of primary selections for a given number of characters is to be kept to some smaller value than this, some restriction must be placed on the freedom of selection of the secondary symbols. Such a restriction is imposed when, in computing the average number of dots per character for a non-uniform code, we take account of the average frequency of occurrence of the various characters in telegraph messages. If this allotted number of dots per character is not to be exceeded in sending a message, the operator must, on the average, refrain from selecting the longer characters more often than their average rate of occurrence. In the language of the present discussion we would say that for certain of the $n_1$ secondary selections the value of $s_2$, the number of secondary symbols, is so reduced that a summation of the information content over all the characters of primary selections involved. This may be written

\[\sum_1^{n_2} log s_2 = n log s\]

where $n$ is the total number of primary symbols or dot lengths assigned to $n_2$ characters. This suggests that the primary symbols furnish the most convenient basis for evaluating information.

The discussion so far has dealt largely with telegraphy. When we attempt to extend this idea to other forms of communication certain generalizations need to be made. In speech, for example, we might assume the primary selections to represent the choice of successive words. On than basis $s$ would represent the number of available words. For the first word of a conversation this would correspond to the number of words in the language. For subsequent selections the number would ordinarily be reduced because subsequent words would have to combine in intelligible fashion with those preceding. Such limitations, however, are limitations of interpretation only and the system would be just as capable of transmitting a communication in which all possible permutations of the words of the language were intelligible. Moreover, a telephone system may be just as capable of transmitting speech in one language as in another. Each word may be spoken in a variety of ways and sung in a still greater variety. This very large amount of information associated with the selection of a single spoken word suggests that the word may better be regarded as a secondary symbol, or sequence of primary symbols. Let us see where this point of view leads us.

The actual physical embodiment of the word consists of an acoustic or electrical disturbance which may be expressed as a magnitude-time function as in Fig. 2, which shows an oscillographic record of a speech sound. Such functions are also typical of other modes of communication, as will be discussed in more detail later. We have then to examine the ability of such a continuous function to convey information. Obviously over any given time interval the magnitude may vary in accordance with an infinite number of such functions. This would mean an infinite number of possible secondary symbols, and hence an infinite amount of information. In practice, however, the information contained is finite for the reason that the sender is unable to control the form of the function with complete accuracy, and any distortion of its form tends to cause it to be confused with some other function.

A continuous curve may be thought of as the limit approached by a curve made up of successive steps, as shown in Fig. 3, when the interval between the step is made infinitesimal. An imperfectly defined curve may then be thought of as one in which the interval between the steps is finite. The steps then represent primary selections. The number of selections in a finite time is finite. Also the change made at each step is to be thought of as limited to one of a finite number of values. This means that the number of available symbols is kept finite. If this were not the case, the curve would be defined with complete exactness at each of the steps, which would mean that an observation made at any one step would offer the possibility of distinguishing among an infinite number of possible values. The following illustration may serve to bring out the relation between the discrete selections and the corresponding continuous
curve. We may think of a bicycle equipped with a peculiar type of steering device which permits the rider to set the front wheel in only a limited number of fixed positions. On such a machine he attempts to ride in such a manner that the front wheel shall follow an irregularly curved line. The accuracy with which he is able to accomplish this will depend upon how far he goes between adjustments of the steering
mechanism and upon the number of positions in which he is able to set it.

By this more or lee artificial device the continuous magnitude-time function as used in telephony is made subject to the same type of treatment as the succession of discrete selections involved in telegraphy.

download videos from youtube and convert to mp4

2010-03-17T09:22:00.000-07:00

Here it a nice tool (easy-to-use script) to download videos from Youtube: youtube-dl.

The new videos (since 2008) are coming using the H.264 video standard, that means you don't need to re-encode those videos into mp4, you just need to change the container, and that is pretty easy to do! :)

ffmpeg -i input.flv -vcodec copy -acodec copy output.mp4

If the video does not come in the H.264 video standard, than you will unfortunately need to re-encode it, and lose quality in the process. :(

Is Reading a Letter-by-Letter Process?

2010-03-04T04:40:00.000-08:00

(William F. Brewer)

(...) In the late 1800s it appears that reading was thought to be a serial processing of the letters making up words and then a serial combination of these words into sentences. However, this position was overthrown by a series of studies coming out of Wundt's laboratory soon after the founding of experimental psychology. Perhaps the most influential work opposed to the letter-by-letter theory was a series of studies carried out by James McK. Cattell [1885a, 1885b, 1886a, 1886b]. Carrell rejected the letter-by-letter (serial) theory of word perception in favor of the whole-word (parallel) approach on the basis of the following kinds of evidence: (a) Words in prose passages can be read almost as fast as lists of letters. (b) The immediate visual apprehension span for letters in prose is much greater than for random letters. (c) Latencies to initiate pronunciation of words are shorter than those for letters. (d) Visual recognition thresholds for words are lower than the thresholds for letters.

For the most part, serious research on the reading process stopped with the rise of behaviorism; however, among those few who continued to work in the area, Cattell's arguments were considered to have shown that word perception is parallel, not serial [Woodworth 1938]. (...)

Misreading: A Search for Causes

2010-02-20T15:07:00.000-08:00

(Donald Shankweiler and Isabelle Y. Liberman)

Because speech is universal and reading is not, we may suppose that the latter is more difficult and less natural. Indeed, we know that a large part of the early education of the school child must be devoted to instruction in reading and that the instruction often fails, even in the most favorable circumstances. [judging from the long history of debate concerning the proper methods of teaching children to read [Mathews 1966], the problem has always been with us. Nor do we appear to have come closer to a solution: we are still a long way from understanding how children learn to read and what has gone wrong when they fail.

Since the child already speaks and understands his language at the time that reading instruction begins, the problem is to discover the major barriers in learning to perceive language by eye. It is clear that the first requirement for reading is that the child be able to segregate the letter segments and identify them with accuracy and speed. Some children undoubtedly do fail to learn to recognize letters and are unable to pass on to succeeding stages of learning to read; but, as we shall see, there are strong reasons for believing that the principal barriers for most children are not at the point of visual identification of letter shapes. There is no general agreement, however, about the succeeding stages of learning to read, their time course, and the nature of their special difficulties. In order to understand reading and compare it with speech, we need to look closely at the kinds of difficulties the child has when he starts to read, that is, his misreadings, and ask how these differ from errors in repeating speech perceived by ear. In this way, we may begin to grasp why the link between alphabet and speech is difficult.

In the extensive literature about reading since the 1890s there have been sporadic surges of interest in the examination of oral reading errors as a means of studying the process of reading acquisition. The history of this topic has been well summarized by Weber [1968] (...).

(...)

We are, in addition, curious to know whether the difficulties in reading are to be found at a visual stage or at a subsequent linguistic stage of the process This requires us to consider the special case of reversal errors, in which optical considerations are, on the face of it, primary. Our inquiry into linguistic aspects of reading errors then leads us to ask which constituents of words tend to be misread, and whether the same ones tend to be misheard. We examine errors with regard to the position of the constituent segments within the word and the linguistic status of the segments in an attempt to produce a coherent account of the possible causes of the error pattern in reading.

(...)

The Word as the Locus of Difficulty in Beginning Reading

One often encounters the claim that there are many children who can read individual words yet do not seem able to comprehend connected text [Anderson and Dearborn 1952; Goodman 1968]. The existence of such children is taken to support the view that methods of instruction that stress spelling-to-sound correspondences and other aspects of decoding are sufficient and many even produce mechanical readers who are expert at decoding but fail to comprehend sentences. (...)

(...)

The Contribution of Visual Factors to the Error Pattern in Beginning
Reading: The Problem of Reversals

We have seen that a number of converging results support the belief that the primary locus of difficulty in beginning reading is the word. But within the word, what is the nature of the difficulty? To what extent are the problems visual and to what extent linguistic?

In considering this question, we asked first whether the problem is in the perception of individual letters. There is considerable agreement that after the first grade, even those children who have made little further progress in learning to read do not have significant difficulty in visual identification of individual letters [Vernon 1960; Shankweiler 1964; Doehring 1968].

REVERSALS AND OPTICAL SHAPE PERCEPTION

The occurrence in the alphabet of reversible letters may present special problems, however. The tendency for young children to confuse letters of similar shape that differ in orientation (such as "b, d, p, q") is well known. Gibson and her colleagues [1962, 1965] have isolated a number of component abilities in letter identification and studied their developmental course by the use of letter-like forms that incorporate basic features of the alphabet. They find that children do not readily distinguish pairs of shapes that are 180-degree transformations (i.e., reversals) of each other at age 5 or 6, but by age 7 or 8, orientation has become a distinctive property of the optical character. It is of interest, therefore, to investigate how much reversible letters contribute to the error pattern of 8-year-old children who are having reading difficulties.

Reversal of the direction of letter sequences (e.g., reading "from" for "form") is another phenomenon that is usually considered to be intrinsically related to orientation reversal. Both types of reversals are often thought to be indicative of a disturbance in the visual directional scan of print in children with reading disability (see Benton [1962] for a comprehensive review of the relevant research). One early investigator considered reversal phenomena to be so central to the problems in reading that he used the term "strephosymbolia" to designate specific reading disability [Orton 1925]. We should ask, then, whether reversals of letter orientation and sequence loom large as obstacles to learning to read. Do they covary in their occurrence, and what is the relative significance of the optical and linguistic components of the problem?

(...)

RELATIONSHIPS BETWEEN REVERSALS AND OTHER TYPES OF ERRORS

It was found that, even among these poor readers, reversals accounted for only a small proportion of the total error, through the list was constructed to provide maximum opportunity for reversals to occur. Separating the two types, we found that sequences reversals accounted for 15 percent of the total errors made, and orientation errors only 10 percent, whereas other consonant errors accounted for 32 percent of the total and vowel errors 43 percent. Moreover, individual differences in reversal tendency were large (rates of sequence reversal ranged from 4 to 19 percent; rates for orientation reversal ranged from 3 to 31 percent). Viewed in terms of opportunities for error, orientation errors occurred less frequently than other consonant errors. Test-retest comparisons showed that whereas other reading errors were rather stable, reversals -- and particularly orientation reversals -- were unstable.

(...)

ORIENTATION REVERSALS AND REVERSALS OF SEQUENCES:
NO COMMON CAUSE?

Having considered the two types of reversals separately, we find no support for assuming that they have a common cause in children with reading problems. Among the poor third-grade readers, sequence reversal and orientation reversal were found to be wholly uncorrelated with each other, whereas vowel and consonant errors correlated 0.73. (...)

ORIENTATION ERRORS: VISUAL OR PHONETIC?

In further pursuing the orientation errors, we examined the nature of the substitutions among the reversible letters "b, d, p, g." Tabulation of these showed that the possibility of generating another letter by a simple 180-degree transformation is indeed a relevant factor in producing the confusions among these letters. This is, of course, in agreement with the conclusion reached by Gibson and her colleagues [1962].

At the same time, other observations [I. Y. Liberman, Shankweiler et al. 1971] indicate that letter reversals may be a symptom and not a cause of reading difficulty. (...)

(...)

Linguistic Aspects of the Error Pattern in Reading and Speech

"In reading research, the deep interest in words as visual displays stands in contrast to the relative neglect of written words as linguistic units represented graphically." [Weber 1968, p.113]

The findings we have discussed in the preceding section suggest that the chief problems the young child encounters in reading words are beyond the stage of visual identification of letters. It therefore seemed profitable to study the error pattern from a linguistic point of view.

(...)

(...) the substantially greater error rate for final consonants than for initial ones is certainly contrary to what would be expected by an analysis of the reading process in terms of sequential probabilities. If the child at the early stages of learning to read were able to utilize the constraints that are built into the language, he would make fewer errors at the end than at the beginning, not more. In fact, what we often see is that the child breaks down after he has gotten the first letter correct and can go no further. We will suggest later why this may happen.

MISHEARING DIFFERS FROM MISREADING

In order to understand the error pattern in reading, it should be instructive to compare it with the pattern of errors generated when isolated monosyllables are presented by ear for oral repetition. (...)

The error pattern for oral repetition shows some striking differences from that in reading. With auditory presentation, errors in oral repetition averaged 7 percent when tabulated by phoneme, as compared with 24 percent in reading, and were about equally distributed between initial and final position, rather than being markedly different. Moreover, contrary to what occurred when the list was read, fewer errors occurred
on vowels than on consonants.

(...)

It is clear from the figure that the perception of speech by reading has problems which are separate and distinct from the problems of perceiving speech by ear. We cannot predict the error rate for a given phoneme in reading from its error rate in listening. If a phoneme were exactly as difficult to read as to hear, the point would fall on the diagonal line that has been dotted in. Vertical distance from the diagonal to any point below it is a measure of the specific difficulty of reading the phoneme as distinguished from listening to it. Although the reliability of the individual points in the array has not been assessed, the trends are unmistakable. The points are very widely scattered for the consonants. As for the vowels, they are seldom misheard but often misread (suggesting, incidentally, that the high error rate on vowels in reading cannot be an artificial transcription difficulties).

(...)

The following analysis illustrates how vowel errors may be analyzed to discover whether, in fact, the error pattern is nonrandom and, if it is, to discover what the major substitutions are. Figure 2 shows a confusion matrix for vowels based on the responses of 11 children at the end of the third grade (Group C2 in Table 6) who are somewhat retarded in reading. Each row in the matrix refers to a vowel phoneme represented in the words (of List 2) and each column contains entries of the transcriptions of the responses given in oral reading. Thus the rows give the frequency distribution for each vowel percentaged against the number of occurrences, which is approximately 25 per vowel per subject.

It may be seen that the errors are not distributed randomly. (...) Hence we may conclude that the error rate on vowels in our list is related to the number of orthographic representations of each vowel.

The data thus support the idea that differences in error rate among vowels reflect differences in their orthographic complexity. Moreover, as we have said, the fact that vowels, in general, map onto sound more complexly than consonants is one reason they tend to be misread more frequently than consonants.

It may be, however, that these orthographic differences among segments are themselves partly rooted in speech. Many data from speech research indicate that vowels are often processed differently from consonants when perceived by ear. A number of experiments have shown that the tendency to categorical perception is greater in the encoded stop consonants than in the unencoded vowels [A. M. Liberman, Cooper et al. 1967; A. M. Liberman 1970]. It may be argued that as a consequence of the continuous nature of their perception, vowels tend to be somewhat indefinite as phonologic entities, as illustrated by the major part they play in variation among dialects and the persistence of allophones within the same geographical locality. By the same reasoning, it could be that the continuous nature of vowel perception is one cause of complex orthography, suggesting that one reason that multiple representations are tolerated may lie very close to speech.

We should also consider the possibility that the error pattern of the vowels reflects not just the complex relation between letter and sound but also confusions that arise as the reader recodes phonetically. There is now a great deal of evidence [Conrad 1964, this volume] that normal readers do, in fact, recode the letters into phonetic units for storage and use in short-term memory. If so, we should expect that vowel errors would represent displacements from the correct vowels to those that are phonetically adjacent and similar, the more so because, as we have just noted, vowel perception is more nearly continuous than categorical. That such displacements did in general occur is indicated in Figure 2 by the fact that the errors tend to lie near the diagonal. More data and, in particular, a more complete selection of items will be required to determine the contribution to vowel errors of orthographic complexity and the confusions of phonetic receding.

Summary and Conclusions

In an attempt to understand the problems encountered by the beginning reader and children who fail to learn, we have investigated the child's misreadings and how they relate to speech. The first question we asked was whether the major barrier to achieving fluency in reading is at the level of connected text or in dealing with individual words. Having concluded from our own findings and the research of others that the word and its components are of primary importance, we then looked more closely at the error patterns in reading words.

Since reading is the perception of language by eye, it seemed important to ask whether the principal difficulties within the word are to be found at a visual stage of the process or at a subsequent linguistic stage. We considered the special case of reversals of letter sequence and orientation in which the properties of visual confusability are, on the face of it, primary. We found that although optical reversibility contributes to the error rate, for the children we have studied it is of secondary importance to linguistic factors. Our investigation of the reversal tendency then led us to consider whether individual differences in reading ability might reflect differences in the degree and kind of functional asymmetries of the cerebral hemisphere. Although the evidence is at this time not clearly supportive of a relation between cerebral ambilaterality and reading disability, it was suggested that new techniques offer an opportunity to explore this relationship more fully in the future.

When we turned to the linguistic aspects of the error pattern in words, we found, as others have, that medial and final segments in the word are more often misread than initial ones and vowels more often than consonants. We then considered why the error pattern in mishearing differed from misreading in both these respects. In regard to segment position, we concluded that children in the early stages of learning to read tend to get the initial segment correct and fail on subsequent ones because the do not have the conscious awareness of phonemic segmentation needed specially in reading but not in speaking and listening.

As for vowels in speech, we suggested, first of all, that they may tend to be heard correctly because they are carried by the strongest portion of the acoustic signal. In reading, the situation is different: alphabetic representation of the vowels possess no such special distinctiveness. Moreover, their embedded placement within the syllable and their orthographic complexity combine to create difficulties in reading.
Evidence for the importance of orthographic complexity was seen in our data by the fact that the differences among vowels in error rate in reading were predictable from the number of orthographic representations of each vowel. However, we also considered the possibility that phonetic confusions may account for a significant portion of vowel errors, and we suggested how this hypothesis might be tested.

We believe that the comparative study of reading and speech is of great importance for understanding how the problems of perceiving language by eye differ from the problems of perceiving it by ear, and for discovering why learning to read, unlike speaking and listening, is a difficult accomplishment.

Speech and Reading

2010-02-12T12:21:00.000-08:00

(R. Conrad)

For a very large number of children, the first word they learn to read is their own name — and one can think of no more meaningful word. Some time may elapse before a second word is confidently added to the reading repertoire. But this first word marks the attainment of a concept of immense intellectual importance: the child accepts the idea that an untidy nonrepresentational pattern of lines in some way symbolizes a name. We say to children, "What does this word say?" In so doing, we intuitively link reading with speaking, dramatically distinguishing between the identification of printed pictures of objects and printed names of objects. Never, in Western culture, would we point to the picture of a dog and ask, "What does this picture say?"

If we examine this intuition in more detail, there is an implication that the printed word says something to the child; the child listens to what is said and then repeats it out loud. But it is not hard to accept that the child, looking at the word, says something to himself, listens to himself, and then repeats what he has heard. In this analysis we are at least within the realm of theoretical possibility. But this does
not make it correct, and in this chapter we shall examine evidence concerning the role of certain processes -- and their interrelationships -- involved in reading.

In no sense is this chapter a broad survey of reading processes. Quite the contrary. We shall be concerned only with the transduction problem of a visual input transformed into a speech-motor output when we read aloud. But the main emphasis will be to consider the speculation that has continuously intrigued students, that the identical transduction necessanly occurs when we read silently to ourselves. We shall consider codes and memories——those systems required for viably maintaining the substance of perceptions during the addition of subsequent perceptions so that higher-order manipulations are possible and textual meaning is achieved. (...)

Speech and Reading in Adults

Does all reading involve speech, and does reading aloud merely add sound? (...) We can
than pose as a primary question, do the processes involved in comprehension occur before the sound is made? In this case, speaking aloud would merely be fulfilling the behavior requirement to read aloud. Or do we make the speech sounds, listen to them, and comprehend what it is we have heard? What we are asking here is whether comprehension of printed material—reading—is possible directly front the visual input.
Or do we have to say words, whether covertly or overtly, in order to understand their meaning? Certainly the latter possibility has a conceptual elegance, since it would fit reading and listening to speech into a single behavioral framework, the only difference being the source of the speech-oneself or another person. But since behavioral mechanisms rarely fall into simple patterns, a discussion of the evidence is necessary; and the reader might just as well be warned at this point that it will
not be conclusive.

(...) At one extreme, the most obvious, we can think of reading with almost full articulation; that is, all articulatory processes are involved except those required for making sounds. Although nothing is heard, lips visibly move to form speech sounds and movements of the speech organs can easily be felt with the fingers. In one sense this is unquestionably silent reading. The continuum then moves toward less directly observable behavior. Lips may be closed and apparently unmoving, movements in
the throat may be too attenuated to be felt. Nevertheless, articulation may still be detected by electromyographic (EMG) and related techniques. Here, the fact that reading is accompanied by electrical activity in muscles required in the production of speech sounds, though no movement is visible to the eye, is taken as evidence for the occurrence of silent articulation of speech during reading. This line of investigation
began with Curtis [1900] and reviews are to be found, for example, in Edfeldt [1960] and McGuigan [1970].

(...) Locke and Fehr [1970] required subjects to read silently (for subsequent recall) two classes of word -- those that did, or did not, include labial phonemes. Using surface electrodes, they also recorded activity of the labial musculature. Though the results were not entirely unambiguous, they seemed to indicate that "labial" words, silently read, do show more movements of the labial muscles than do "nonlabial" words. At any rate the procedure is novel and highly promising.

But even the complete absence of detectable speech-motor activity does note preclude the occurrence of silent speech in the form of speech imagery. We are not saying here that this ever happens, only that it can happen. If auditory imagery is a genuine biological phenomenon, then the sounds of speech must be included in its definition, and silent reading can be accompanied by a succession of auditory speech images that might have the same psychological function in the reading process as does silent articulation. Granted auditory imagery in this context, then speech-motor imagery, even though unresponsive to EMG recording, must be included in this definition as well. If in imagination we can move a leg, then there is no reason why in imagination we should not be able to move a tongue and so image the feel of silent speech without articulation and without imaged sound. All these phenomena would be silent speech. The fact that no visible lip movements is observed would not preclude the presence of silent speech in reading. EMG silence equally does not preclude silent speech. As always, it is easier to prove the presence of a phenomenon than its absence.

(...)

[McGuigan 1970] uses EMG recording. As we have said, this category has a history going back to the beginning of this century; and almost without exception, results show more EMG response from speech muscles during silent reading than during rest. One of the most comprehensive and carefully controlled studies is that of Edfeldt [1960], who reported articulation during silent reading with almost all of 84 subjects. (Pervesely, Edfeldt does not report whether any of these subjects gave negative introspections). There are indeed so many good studies reporting the presence of articulation during silent reading, that we might be justified in concluding that the case is proved. But of course, the case that appears to be proved is that silent reading is accompanied by articulation. What is far from proved is that articulation is necessarily involved in silent reading. This kind of imperative seems most unlikely. No one has convincingly shown comprehension to be seriously impaired directly as a result of preventing articulation in some way. This latter is not easily accomplished by mechanical means [Novikova 1966]. Indeed, the contrary may be true.
Hardyck, Petrinovich et al. [1967], using a conditioning procedure, reported the complete inhibition of articulation during silent reading with unimpaired comprehension. A later report qualified this [Hardyck and Petrinovich 1970]. In any case, absence of articulation does not preclude the use of speech imagery, so that comprehension when there is no EMG response could still be based, as a requirement, on silent speech.

Reading without Speech

Since this paper is directly concerned with the role of speech in reading and in learning to read, it is certainly pertinent to consider the nature and the effectiveness of reading skill in the deaf, who have had no experience of aural speech. There are of course, other pathologically handicapped populations who do not speak, such as certain aphasics. But unless deafness is present, opportunity to hear speech has been present. We have seen how tenaciously the normal hearing person clings to his phonological code when he reads material that he is to remember; and we have seen the way young children appear spontaneously to come to use ts phonological name-code when they memorize a series of pictures. When, through experimental manipulation, the phonological code becomes difficult to use, STM is gravely impaired. We have therefore suggested that there is a close association between silent reading and silently speaking because comprehension requires memory. The STM involved seems best supported by phonological coding.

Only among the deaf can we find people with no speech experience -- or at any rate with relatively little -- who for our present purposes therefore provide an invaluable control. We can simply ask the question: can the deaf learn to read? Were the answer an emphatic negative much of our enquiry would come to a comfortable end, since we would be very close to proving that reading is possible only when phonological coding is available. But of course the truth is not as simple as that, and most deaf children do, to some extent, learn to read. That immediately tells us that for hearing people, phonological coding is a preference, not a necessity. Knowing the versatility of man, this is what we would have expected. But since it has turned out to be exceedingly difficult to get hearing subjects to abstain from their predilection for phonological coding, we might hope that studies of the deaf would give us clearer insight into the rules governing the development and use of reading codes in general. By this we refer to the kinds of transductions that occur between seeing a printed item and storing it in a form in which it is available for future use, so that we can say: an item is remembered; a phrase is comprehended.

The paradox of a "normal deaf" population is too great for the term to be meaningful. In the first place, anyone whose hearing does not fall within defined "normal" limits is deaf. In the studies to be discussed, we are concerned with a category of profound deafness. Over the range of useful speech frequencies between 250 and 4000 Hz, our subjects, mostly children, had hearing losses of not less than about 75 dB in their better ear. These children, even with hearing-aid amplification, would have very little awareness of different speech sounds. Unless they can see lips moving, they would not know that someone in a room was speaking. They are profoundly deaf. Second their medical records would show that they were either born deaf -- sometimes of deaf parents -- or became profoundly deaf within the first year or so of life; they have never used normal speech. These are features common to any person who can loosely be described as "congenitally and profoundly deaf". All of the deaf subjects of the studies to be discussed are in this category. But by the time a deaf child is likely to take part in reading experiments, variations in home background, intelligence, other pathology, and particularly in the educational theories that have guided his school training, will become important sources of experimental "noise." In using these subjects in experiments we have to confine ourselves to relatively simple questions and be content only with large experimental effects.

Probably in most schools for the deaf some attempt is made to teach deaf children to speak. There are some schools where no other mode of communication is used, come what may, between teachers and pupils. Other schools regard communication by no matter what means, as essential. Even less then, than with normal children, can we talk of deaf children "on average." But no matter how speech oriented a school might be, children who are profoundly deaf from an early age exceedingly rarely attain a quality of speech that can be readily understood by Strangers. It is exceedingly rare that a hearing person can address such a deaf person normally, even though making sure that his lips can be seen, and get normal comprehension of his speech at normal speeds. It is exceedingly rare to see two such deaf children speaking to each other using the language that hearing persons would use. All of these things do occur, but they are levels of speech and language skills that certainly fewer than one in a hundred profoundly deaf children achieve.

If then speech is a skill acquired by the deaf with immense difficulty when acquired at all, from all that has gone before we would expect reading also to present grave difficulty to the deaf. Apart from anything else, this would be due to severe vocabulary deficiency and to the handicap of never having acquired the easy use to the innumerable rules of grammar that the rest of us pick up through hearing before we usually begin to learn to read. But there is something else also. There is the fact of only partial, or even of complete absence of, availability of phonological coding as an aid to comprehension. And indeed there is very good evidence that profoundly deaf children have great difficulty in learning to read. Myklebust [1960] reports on a number of studies of reading ability in the deaf.

For example, on the Columbia Vocabulary Test, at 9 years the mean score for normal children is about 20; for deaf children it is just over 3. At 11 years the respective values are 33:6; at 13 years, 43:10; at 15 years, 63:11. Not only is the difference huge, but it increases with age. The 15-year-old deaf child has a much poorer vocabulary than a 9-year-old normal child. In every aspect of the grammatical handling of words the deaf child is many years behind the hearing child on attainment tests. It is not therefore surprising that similar discrepancies are reported when memory span for verbal material is examined. Blair [1957], Furth [1964], Olsson and Furth [1966], and others have shown that deaf children are grossly inferior in digit span to hearing children. But when the material to be remembered consists of shapes that do not readily have names, deaf and hearing children have the same span. Furthermore the Olsson and Furth data show that whereas for hearing subjects the difference in span between digits and shapes is substantial, for the deaf it is quite small. It is not impossible then, that for the deaf, digits are just another set of shapes. There is no internal evidence in the Olsson and Furth study to indicate whether the deaf group did in fact code digits by shape. Olsson and Furth suggest that they could have finger—spelled them; there is uncertainty. The point is that there is some suggestive evidence here that phonological coding has unique advantages for STM over certain other codes, and among these we must include visual codes. This is unquestionably a loose argument since we do not know for certain what kind of codes in this experiment either group used with any of the test material. Then, as Furth [1966] points out elsewhere, any direct comparison of cognitive abilities between deaf and hearing must be taken with great caution because of innumerable differences that cannot be taken into account when matching groups.

Since many studies show deaf and hearing to have similar span on material not easily verbalized, but inferiority of the deaf when nameable items are used, the need is emphasized for a clearer understanding of the STM code used by the deaf when presented with words and other items having familiar names. We have continuously pointed out the important role that phonological coding has in reading: but we are equally clear that the deaf do read. They could be using a phonological code very ineffectively, or they could be using some other code or codes that are in themselves less efficient for the purpose than is a phonological code.

(...)

Now we have tended to discuss codes used by the deaf when they read printed verbal material as if we believe that when an articulatory code is not used, then a visual code is used. This is not our belief. It so happens that a good deal of the data we have used in this discussion can be taken to support this view. But this is largely because tasks have been designed that would favor visual coding were it available to the deaf. We are quite sure that other codes are also available to most of the deaf that are based on finger-spelling, signing, and lip-reading at the least. On present knowledge it is much harder to construct reading test material that could greatly benefit, for example, a finger-spelling code. Indeed it seems plausible that with reading skills, since the deaf do not have available what seem to be the most efficient codes, they could very well make extensive use of multiple coding, no one code being particularly efficient, to a much greater extent than do those of us with normal sensory experience.

(...)

By analogy, words have been "designed" to provide high auditory discriminability. When we read we continue to take advantage of this fact. The visual appearance of words has limited relevance. The deaf try to read a language perfectly adapted for the use of hearing people. They are faced with many commonly used words of the same length and general configuration; the majority of words begin with a minority of different consonants, usually followed by a vowel, all of which sound different but which in print, except for i, look quite similar. Just as we try to teach the deaf a spoken language that permits them to use only minimally those cognitive abilities that are intact, so too we try to teach them a written language, making the tacit assumption that when a deaf child sees a printed word he will cognitively “do” with it just what a hearing child will.

In summary, what this section has tried to do is to show that speech or speech-based codes are not necessary for reading. We have agreed that deaf children who do not use speech-based codes are poor readers. But we have argued that the reasons for this are far from simple. In particular it may be that nonspeech codes, when developed by training to the extent that speech codes are, would not in themselves be inefficient. But when used to transduce printed words that derive from spoken forms, then they are bound to be at a disadvantage.

The legacy of Zellig Harris: Language and information into the 21st century

2010-02-11T11:33:00.000-08:00

(John Goldsmith)

Zellig Harris (1909–1992) cast a long shadow across twentieth century linguistics. In mid-century, he was a leading figure in American linguistics, serving as president of the Linguistic Society of America in 1955, just a year before Roman Jakobson. It is fair to say that during that decade—the years just before generative grammar came on the scene—Zellig Harris and Charles Hockett were the two leading figures in the development of American linguistic theory. Today, I daresay Harris is remembered by most linguists as the mentor and advisor to Noam Chomsky at the University of Pennsylvania—and the originator of transformational analysis.

(...)

Harris’s work must be situated in terms of the conflict between two visions of linguistic science: the MEDIATIONALIST view, which sees the goal of linguistic research as the discovery of the way in which natural languages link form and meaning, and the DISTRIBUTIONALIST view, which sees the goal as the fully explicit rendering of how the individual pieces of language (phoneme, syllable, morpheme, word, construction, etc.) connect to one another in the ways that define each individual language. The mediationalist view lurks behind most conceptions of language study, formal and nonformal, but it was Harris’s view that each successive improvement in linguistic theory took us a step further AWAY from the mediationalist view, much as advances in biology led scientists to understand that the study of living cells required no new forms of energy, structure, or organization in addition to those which were required to understand nonliving matter. Harris had no use for mediationalist conceptions of linguistics. For linguists in 2005, steeped as we are in an atmosphere of linguistic mediationalism, this makes Harris quite difficult to understand at first.

Harris’s goal was to show that all that was worthwhile in linguistic analysis could best be understood in terms of distribution of components at different hierarchical levels, because he understood—or at least he believed—that there was no other basis on which to establish a coherent and general linguistic theory. His genius lay in the construction of a conception of how such a vision could be put into place concretely.

Harris’s view, from his earliest work through his final statements in the early 1990s, was that the best foundational chances for linguistics were to be found in stablishing
a science of EXTERNAL LINGUISTIC FACTS (such as corpora, though they would typically be augmented by other external facts, like speaker judgments), rather than a science of internalized speaker knowledge. (...)

(...)

Harris did not appear to make a great effort to make his conclusions easily accessible to the reader. And yet once his ideas are understood, it is hard to deny that his way of stating them is direct, elegant, and striking. Let us approach the central idea of all of Harris’s work, as summarized by Harris himself in his introductory paper to this volume:

"The structure of language can be found only from the non-equiprobability of combination of parts. This means that the description of a language is the description of contributory departures from equiprobability, and the least statement of such contributions (constraints) that is adequate to describe the sentences and discourses of the language is the most revealing."

Picking this apart into pieces:

1. Linguistic analysis consists of building a representation out of a finite number of
formal objects.

2. The essence of any given language is the restrictions, or constraints, that it places on how the pieces may be put together—these may be phonemes, morphemes, constituents, what have you. If there were no structure, then pieces could be put together any which way; structure MEANS — it is nothing more or less than — restrictions on how pieces can be put together.

3. These restrictions may be absolute ('no pk' clusters are permitted in this language’) or, much more likely, they are statements of distribution, best expressed in the mathematics of probability. A crude reformulation of this would be in the language of markedness, which is arguably an informal way of talking about distributional frequencies. A better way is to use the mathematical vocabulary of distributions, which is to say, probability theory.

4. A formal system can be described formally in a multitude of ways. These are not equivalent: there is a priority among them based on their formal length. In general, one will be significantly shorter than the others, and knowing its length is important.

It is probably impossible to understand the intellectual pull of this research program if one does not appreciate the revolutionary character (and the perceived success) of the phoneme. If the phoneme today seems passe, the discarded error of an earlier generation, then today’s linguist should think of its descendant—for most of us, the
idea of an underlying segment. (...)

(...)

It was his view that the important relationship between sounds lay not in their phonetics, but in their DISTRIBUTION (...). What tells us that the flap and the other t’s of English are realizations of a single phoneme /t/ is not the similarity of sound, but the complementarity and predictability of the distribution.

"It is pointless to mix phonetic and distributional contrasts. If phonemes which are phonetically similar are also similar in their distribution, that is a result which must be independently proved. For the crux of the matter is that phonetic and distributional contrasts are methodologically different, and that only distributional contrasts are relevant while phonetic contrasts are irrelevant.

This becomes clear as soon as we consider what is the scientific operation of working out the phonemic pattern. For phonemes are in the first instance determined on the basis of distribution. Two positional variants may be considered one phoneme if they are in complementary distribution; never otherwise. In identical environment (distribution) two sounds are assigned to two phonemes if their difference distinguishes one morpheme from another; in complementary distribution this test cannot be applied. The distributional analysis is simply the unfolding of the criterion used for the original classification. If it yields a patterned arrangement of phonemes, that is an interesting result for linguistic structure."

(...)

Rudolf Carnap had (...), in The logical syntax of language (published in 1934 under the title Logische Syntax der Sprache), argued for a coming together of formal syntax and formal logic: by formal, he meant analysis ignoring meaning and considering only categories and combinations of symbols; by syntax, the rules by which items are combined to form expressions (sentences); and by logic, the rules by which valid inferences from one sentence to another can be made. The contrast between syntax and logic was dubbed by Carnap (in English) as the difference between FORMATION rules and TRANSFORMATION rules, an interesting terminological suggestion and one that may have later influenced Harris.

Component Processes in Reading: A Performance Analysis

2010-02-10T11:36:00.000-08:00

(Michael I. Posner, Joe L. Lewis, and Carol Conrad)

Introduction

A detailed analysis of the internal structures and mental operations involved in the process of reading might help us understand problems in acquiring the skill. This is a point that our keynote speaker [Gibson 1965] made several years ago. In the last few years there has been a considerable advance in the development of techniques used to isolate stages of processing and their interrelationships [Neisser 1967; Posner 1969; Sternberg, 1969]. This paper is an effort to review both the techniques and the results that might aid in elucidating the component processes in reading.

(...)

Isolable Subsystems

Let us be quite concrete. A young child who has never before seen the symbol "A" must be aware primarily of the visual form of the letter. But this is not so with an adult. Consciousness of the letter is suffused with past experience: its association to other visual forms (e.g., "a”), the phoneme /a/, its status as a vowel, and as the first letter of a list called the alphabet. Yet, even in the skilled reader, by appropriate
experimental technique, we can isolate the visual system processes from these other influences. We can, in fact, argue that the visual processes represent in the adult an isolable subsystem the properties of which can be studied.

Experiments showing that the visual process is an isolable subsystem of letter processing in the adult [Posner 1969] suggest that there are important psychological problems involved in passing from one subsystem to another (e.g., visual to name). Perhaps it is at the boundaries between any two isolable subsystems that special difficulties in cognitive processing lie. Indeed, the problem of coordinating modality-specific subsystems may represent one explanation of the difliculty in the seemingly simple translation from a visual word to the word name.

The idea of an isolable subsystem is a complex one [Miller 1970]. In the recent experimental literature there have been many eFforts to discover serial “stages" of processing [Clark and Chase 1971; E. E. Smith 1968; Sternberg 1959, Trabasso 1970]. These tasks tend to be ones in which one stage must depend directly upon the outcome of the prexious stage. There is still dispute about the details of these models (eg., whether the comparison stage is serial), but they have had sufficient success to show that internal mental operations can be isolated for study.

(...)

Visual Codes

The words that you are presently reading are unique conngurations of print. The names that these words represent are abstractions in the sense that they stand for a variety of perceptually ditcferent visual forrns (eg., "PLANT, plant”) and auditory patterns [eg., the word plant spoken by a male or a female). The name of a word gains its meaning from the semantic structure to which it is related. The word plant may be related to a structure dealing with living things, or to one dealing with labor unions and assembly lines [Quillian 1969]. At one level the word is a visual code, at another a name, and at still another, an aspect of the overall semantic structure of which it is a part.

AN EXPERIMENTAL METHOD

The mental operations that transform one code into another can be observed in the time required for making classifications. Suppose that the subject is shown a pair of items and is asked to press, as quickly as possible, one key if they are "same" and another if they are "different." Figure 1 (left diagram) illustrates the results from an experiment in which items were letters and the definition of “same" was "both vowels” or "both consonants” [Posner and Mitchell 1967]. If the letters were identical in physical form ("A A"), the reaction time was faster than if they had only a name in common (e.g., "A a"), which in turn was faster than when items shared only the same class (e.g., “A e,” both vowels). A similar result [Schaeffer and Beller 1970] is shown (right diagram) for an experiment in which word pairs were used and the subject was required to press the key when both words were "living things" or both were "nonliving things". These figures illustrate the method measuring mental operations by the amount of time they require.

ISOLABILITY

Can we isolate operations that are performed on visual information? The problem is how to determine if the operations performed on letters or words use visual representations rather than letter names or semantic information. Suppose that a pair of items is presented simultaneously. The subject is then required to press one key if the two items have the same name and another key if the names are different. If the two are physically identical (e.g. "A A") it is logically possible to base the match upon a visual form. On the other hand, for letters like "A a", which are not similar in physical form, the match is more likely to be based upon a learned correspondence between the visual forms such as the letter name.

Experimental data show that these logical distinctions apply to actual performances of subjects. The time for matching identical letters (e.g., "A A") is faster than that for upper and lower case forms of the same letters (e.g. "A a"). (...)

(...)

FAMILIARITY

The isolability of visual and name processes, even for letters, allows us to study properties of the visual code independently of names. For example, it is now clear that any number of letters may be matched simultaneously, as long as they are physically identical. This means that the contact between external letters and their internal pattern recognizers can go in parallel and with no interference [Beller 1970; Donderi and Case 1970]. (...)

In her 1965 paper, Gibson argues that the primary units of analysis for reading are spelling patterns rather than single letters. It is possible now to show that these familiarity effects occur within the visual code and do not depend upon feedback from the letter names. Recent experiments have allowed us to study the influence of past experience upon mental operations within the visual system. If a subject is required to match two strings of letters to determine if they are physically identical, he can do so much faster if they form a familiar word than if they are nonsense strings [Eichelman 1970; Krueger 1970]. (...)

Moreover, the word can appears to form a unit. If the subject has to identify a single letter from a brief exposure, he can do so as efficiently when a words is presented as when a single letter is presented [Reicher 1969; Wheeler 1970]. It thus appears that having had past experience with certain sequences of letters allows us to perform matching and other visual operations upon them with great efficiency.

(...)

(...) In one study [Smith, Lott el al. 1969] it was found that subjects scanned a visual array just as rapidly whether the visual patterns were familiar ("PLANT, plant") or quite new ("pLaNt"), provided that the letters were equated for visibility. (...)

Nothing we have said suggests a solution to the more general problem of pattern recognition. We simply do not know how the input is brought into contact with the internal system. (...)

(...)

Name Codes

The concept of the name of a word or letter is an extremely important one for the psychology of reading. Many theories implicitly assume that the internal representation that stands for the name of a word is the same regardless of the modality through which the information was received [Morton 1969]. This assumption greatly simplifies an analysis of reading. The unique problem of reading would then involve mainly converting from a visual to a name code. From there on, comprehension would be based on mechanisms already present for listening. We have already reviewed one objection to this idea, namely, the view that meaning is connected directly to the visual forms. A second objection is that subjects can recall the channel of entry by which a stimulus was presented [Murdock 1967]. This objection, however, can be met by recognizing that the activation of a name code does not obliterate information about the past history of the input [Posner 1969].

Name Codes and Meaning

It is obvious that knowing the name of a word is not the same as knowing its meaning. James [1890] commented on his point (Vol.I, p. 263), "it is more difficult to ascend to the meaning of word than to pass from one word to another; or to put it otherwise, it is harder to be a thinker than to be a rhetorician, and on the whole nothing is commoner than trains of words not understood." It is well known that the word associations of children often involve similarity of word sound, a type of association that is reduced in frequency later in life. If subjects are asked to signify that they have read a word, they respond much more quickly than when required to signify that they understand it [Wickens 1970]. (...)

(...)

Reading, the Linguistic Process, and Linguistic Awareness

2010-02-09T09:33:00.000-08:00

(Ignatius G. Mattingly)

More recently, however, the perception of speech has come to be regarded by many as an “active" process basically similar to speech production. The listener understands what is said through it process of “analysis by synthesis" [Stevens and Halle 1967]. Parallel proposals have accordingly been made for reading. Thus Hochberg and Brook [1970] suggest that once the reader can visually discriminate letters and letter groups and has mastered the phoneme-grapheme correspondences of his writing system, he uses the same hypothesis-testing procedure in reading as he does in listening (Goodman’s [1970] view of reading as a "psycholinguistic guessing game" is a similar proposal). Though the model of linguistic processing is different from that of Bloomfield and Fries, the assumption of a simple parallel between reading and listening remains, and the only differences mentioned are those assignable to modality, for example, the use which the reader makes of peripheral vision, which has no analog in listening.

(...)

We know that all living languages are spoken languages, and that every normal child gains the ability to understand his native speech as part of a maturational process of language acquisition. In fact we must suppose that, as a prerequisite for language acquisition, the child has some kind of innate capability to perceive speech, In order to extract from the utterances of others the "primary linguistic data" that he needs for acquisition, he must have a "technique for representing input signa1s" [Chomsky 1965, p. 30].

In contrast, relatively few languages are written languages. In general, children must be deliberately taught to read and write, and despite this teaching, many of them fail to learn. Someone who has been unable to acquire language by listening -- a congenitally deaf child, for instance -- will hardly be able to acquire it through reading; on the contrary, as Liberman and Furth [Kavanagh 1968] point out, a child with a language deficit owing to deafness will have great difficulty learning
to read properly.

The apparent naturalness of listening does not mean that it is in all respects a more efficient process. Though many people find reading difficult, there are a few readers who are very proficient: in fact, they read at rates well over 2000 words per minute with complete comprehension. Listening is always a slower process: even when speech is artificially speeded up in a way which preserver frequency relationships, 400 words
per minute is about the maximum possible rate [Orr, Friedman et at. 1965]. It has often been suggested [e.g., Bever and Bower 1966; Bower, 1970] that high-speed readers are somehow able to go directly to a deep level of language, omitting the intermediate stages of processing to which other readers and all listeners must presumably have recourse.

(...) The listener is processing a complex acoustic signal in which the speech cues that constitute significant linguistic data are buried. Before he can use these cues, the listener has to "demodulate" the signal: that is, he has to separate the cues from
the irrelevant detail. The complexity of this task is indicated by the fact that no scheme for speech recognition by machine has yet been devised that can perform it properly. The demodulation is largely unconscious; as a rule, a listener is unable to perceive the actual acoustic form of the event which serves as a cue unless it is artificially excised from its speech context [Mattingly, Liberman et al. 1971]. The cues are not discrete events well separated in time or frequency; they blend into one another; we cannot, for instance, realistically identify a certain instant as the ending of a formant transition for an initial consonant and the beginning of the steady state of the following vowel.

The reader, on the other hand, is processing a series of symbols that are quite simply related to the physical medium that conveys them. The task of demodulation is straightforward: the marks in black ink are information; the white paper is background. The reader has no particular difficulty in seeing the letters as visual shapes if he wants to. In printed text, the symbols are discrete units. In cursive writing, of course, one can slur together the symbols to a surprising degree without loss of legibility. But though they are deformed, the cursive symbols remain essentially discrete. It makes sense to view cursive writing as a string of separate symbols connected together for practical (convenience; it makes no sense at all to view the speech signal in this way.

(...)

Reading as a Language-Based Skill

Our view is that reading is a language-based skill like Pig Latin or versification and not a form of primary linguistic activity analogous to listening. From this viewpoint, let us try to give an account, necessarily much oversimplified, of the process of reading a sentence.

The reader first forms a preliminary, quasiphonological representation of the sentence based on his visual perception of the written text. The form in which this text presents itself is determined not by the actual linguistic information conveyed by the sentence but by the writer's linguistic awareness of the process of synthesizing the sentence, an awareness which the writer wishes to impart to the reader.

(...)

How can we explain the very high speeds at which some people read? To say that such readers go directly to a semantic representation, omitting most of the process of linguistic synthesis, is to hypothesize a special type of reader who differs from other readers in the nature of his primary linguistic activity, and differs in a way which we have no other grounds for supposing possible. As far as I know, no one has suggested that high-speed readers can listen, rapidly or slowly, in the way they are presumed to read. A more plausible explanation is that linguistic synthesis takes place much faster than has been supposed, and that the rapid reader has learned how to take advantage of this. The relevant experiments (summarized by Neisser [1967]) have measured the rate at which rapidly articulated or artificially speeded speech can be comprehended, and the rate at which a subject can count silently, that is, the rate
of "inner speech". But since temporal relationships in speech can only withstand so much distortion, speeded speech experiments may merely reflect limitations on the rate of input. The counting experiment not only used unrealistic material but assumed that inner speech is an essential concomitant of linguistic synthesis.

Rate Distortion Theory

2010-02-05T12:26:00.000-08:00

(Elements of Information Theory, Thomas M. Cover and Joy A. Thomas)

The description of an arbitrary real number requires an infinite number of bits, so a finite representation of a continuous random variable can never be perfect. How well can we do? To frame the question appropriately, it is necessary to define the “goodness” of a representation of a source. This is accomplished by defining a distortion measure which is a measure of distance between the random variable and its representation. The basic problem in rate distortion theory can then be stated as follows: Given a source distribution and a distortion measure, what is the minimum expected distortion achievable at a particular rate? Or, equivalently, what is the minimum rate description required to achieve a particular distortion?

One of the most intriguing aspects of this theory is that joint descriptions are more efficient than individual descriptions. It is simpler to describe an elephant and a chicken with one description than to describe each alone. This is true even for independent random variables. It is simpler to describe X1 and X2 together (at a given distortion for each) than to describe each by itself. Why don’t independent problems have independent solutions? The answer is found in the geometry. Apparently, rectangular grid points (arising from independent descriptions) do not fill up the space efficiently.

Fórmulas Matemáticas no Blogger

2010-02-05T09:50:00.000-08:00

Para adicionar fórmulas ao seu Blogger é bem fácil. Siga o link:

Veja aqui um primeiro teste
$A = \sum_{i=0}^{N} x_i$

Written Signs

2010-02-02T08:28:00.002-08:00

(Writing Systems: An introduction to their linguistic analysis, Florian Coulmas)

These technical developments have repercussions on the structure of the signs and the way they are processed. Recognition of the signs is no longer based on similarity but on discrimination, as a pictorial likeness is gradually replaced by the necessity to distinguish one sign from another. Differentiation thus becomes the principal design feature of the signs. For example, that sign of a bull resembles a bull is now less important than that it differs from the sign of a cow.

(...)

The relationship between signs and objects is superseded by multiple relationships between signs and other signs as the scribes' chief concern. The signs thus become part of a graphic system characterized by negative differentiation. The underlying principle is that the many signs are to be kept from becoming confused with one another, much like the units of a language. The creation of new signs follows the same principle when lines are added to existing signs or one sign is adjoined to another. Contrast with all other signs become a defining feature of every sign.

Writing Auto-indexicality

2010-02-02T08:28:00.001-08:00

Every written document not only embodies the message 'I am meant to be read' but also instructions, however indirect, as how this can be done. In other words, the systematic make-up of writing contains a key to its own decipherment.

(Writing Systems: An Introduction to Their Linguistic Analysis, Florian Coulmas :21)

Vowel Incorporation

2010-02-02T08:27:00.001-08:00

Most writing systems are interpreted as referring in some way to the phonetic composition of speech forms. In the process the natural continuous flow of sound is artificially broken up into discrete units of various size. The syllable is an intuitively salient unit exploited to this end by several ancient and modern writing system such as Assyrian cuneiform, Cypriot and Japanese kana (...). Syllables are typically composed of consonants and vowels, which, in the Western tradition, as a reflection of the Greek alphabet are both uniformly considered sound segments, while in Semitic writing consonants and vowels are conceptualized and symbolized differently. The use of 'matres lectionis' in archaic Semitic documents is clear evidence that the Semitic scribes had a notion of a vowel as a unit of language. For reasons having to do with the conservative nature of writing systems in general and with the semantic significance of consonants in Semitic languages, they chose not to, or were not able to, treat both classes of sounds in the same manner.

Linguistic analysis

All writing systems are based on, and hence more or less explicitly incorporate a linguistic analysis. In the case of Indian writing systems this is especially obvious.

Symbols as a Generalized Language

2010-02-02T08:22:00.000-08:00

Writing is an example of a language isomorph in that it has a close part-to-part correspondence with natural language and scientific formulae are extensions of language in that they begin with language and then go on to constructions which are language-like, but not actually used in natural language. Symbols are still wider generalizations of language than either isomorphs or extensions. In the widest sense a symbol is anything, linguistic or non-linguistic, which stands for or "symbolizes", something else. The symbol "a" stands for the sound [a], the visual symbol "|", whatever you call it, stands for the number 'one' in more than one system of writing. A repeated low-pitch horn may stand for a warning that there is a heavy fog in the harbour. A system however has to be something which cann be conveniently produced, presented, and perceived without necessary perceiving the object it stands for.
(...)
we shall follow the therminology of Charles W. Morris (...) Moris deals with is not symbols, but, as the title of his work implies, signs, of which symbols form a special case. For example, lowering clouds are a sign of rain, a shiny wet road is a sign that the road is slippery, but the road sign which says "Slippery When Wet" is not only a
sign, but also a symbol. In general, as we have noted above, a symbol is something which can be conveniently produced and has a conventionalized, usually arbitrary, relation to what is symbolized.

what is one symbol?

In analysing a complexity thing, such as symbolic systems, there is always the twofold problem of (1) identification or differentiation on the one hand, and (2) idividualization or segmentation on the other. We have already met with similar problems in the case of phonemes and words. More generally, we can ask: What similar of different things can be classed together as instances of the same symbol? This is a question of kind. Or we can ask: How much a chunk of a thing extending in space or time or both in space and time (such as gestures) shall be considered one piece of a symbol? This is a problem of size. To revert to our linguistic interest in things grammatic, the former is a paradigmatic problem, while the latter is a syntagmatic problem. This is in fact not too far from Morris's terminology, since he calls the study of the structure of signs (including symbols) themselves syntactics, which of course has a much wider application that syntax in the grammatical sense.

(1) To take the problem of identity of symbols first, it will be more convenient to regard a symbol, not as one event or one thing, but as a collection of events or things considered as members of a class, in other words, a symbol is usually taken as a type rather than a token. On the other hand one instance of a symbol, or token, such as an utterance made on one occasion, is often termed a 'signal'. In common usage, one speaks of signals usually in connection with special forms of visual and other forms of communication other that linguistic forms, but there is no reason why a signal in the sense of one instance of the use of a symbol should not include language.
(...)

(2) As to the problem of segmentation, there are two sides to consider: (a) What is one symbol and what is a complex of symbols? (b) Where does a particular symbol begin and end? These are obviously generalizations of corresponding linguistic problems of subunits of language, with which we are already familiar. As for the complexity of symbols, no upper limit can be set. As Rudolph Carnap has noted, to any sentence which is reputed to be the longest sentence possible, one can always add the co-ordinate clause 'and the moon is round', which makes it a longer sentence. (...) The lower limit to the size of a symbol is not the smallest physical element which is perceivable, but a symbol which, even if perceivable when subdivided, whould no longer be a symbol (or a set of symbols) in the system of which it is a part.

(...)

Symbol and object

(1) Symbols and icons. The normal relation between a symbol and its object, or denotatum in Morris's terminology, is conventional, arbitary, and fortuitous. There is usually no similarity or casual relation between the two. There is for example nothing intrinsically long about the English word 'long' or intrinsically short about the word 'short'. In fact the word 'short' is longer not only graphically but also phonetically and foreigners often tend to pronounce it 'shot' in order so make it sound more symbolic -- symbolic in the popular sense we noted above: 'fitting, expressive, consonant, appropriate', which is precisely the opposite of 'conventional, arbitrary', etc. In this popular sense red is symbolic of danger, stop, etc., because it is physiologically more impressive.

(...)

4. Ambiguity, vagueness, and generality. Symbol and object may correspond in the relation of one to one, one to many, many to one, or many to many, understanding of course that one symbol may consist of a class of various members whose differences do not matter.

Syllables

2010-02-02T08:20:00.000-08:00

(Writing Systems: An introduction to their linguistic analysis, Florian Coulmas)

Intuitive notions of the syllable are vague. Attempts at precision move the discussion to a different level of analytical notions defined in a theoretically justified way. In phonology, the syllable is seen either as the minimum unit of sequential speech sounds or as a unit of the metrical system of a language. Certain theories consider the syllable as a basic phonological unit 'sui generis', while others derive its properties from those of the composite phonemes. Clearly, a syllable is a unit of articulation, and although a universally accepted articulatory definition is not available, phoneticians of different schools are agreed that syllables possess psychological reality for speakers. A syllable is a unit of speech that can be articulated in isolation and bear a single degree of stress, as in English, or a singles tone, as in Chinese. Different languages allow for different syllables. The specific structure of possible syllables is thus part of the phonological system of a language. In very general terms, syllables are units of speech consisting of an obligatory nucleus, usually a vowel (V), and optional initial and final margins, usually consonants (C). An alternative way of describing the structure of the syllable is to divide it into onset and rhyme, where the onset is the initial margin and the rhyme is further subdivided into peak and coda. A syllable with a vowel in coda position is called 'open', and a syllable with a consonant in coda position 'closed'.

(...)

The syllable is also the domain of stress, another feature of cross-linguistic variation. In French, stress is not very important, it rarely affect meaning. In English it can be distinctive, as in 'increase with stress on the first syllable, a noun, and in'crease, a verb, stressed on the second syllable. (...)

In some languages vowel length is distinctive, which means that there are minimal pair of syllables that differ in phonological time only. (...)

The syllable further function as the unit to which a pitch level is assigned. Languages that use pitch level to distinguish words are known as 'tone languages', and distinctive pitch levels are called 'tones'. In tone languages it is relations between the pitch of different syllables rather than the absolute pitch that is important. (...)

To summarize, segmental composition, stress, duration and tone are properties of the syllable. The importance of these features varies across languages and, although the syllable is crucial as a unit within which the distribution of phonological features can be stated, it is best defined as a unit for each language separately. This has important consequences for the analysis of syllabic writing systems.

Subliminal Perception

2010-02-02T08:19:00.001-08:00

(Cognitive Psychology: A Student's Handbook, Michael W. Eysenck and Mark T. Keane)

In 1957, it was reported in the press that James Vicary flashed the words EAT POPCORN and DRINK COCA-COLA for 1/300th of a second numerous times during the cinema showing of a film. This subliminal advertising allegedly led to an 18% increase in the cinema sales of Coca-Cola and 58% increase in popcorn sales. However, the film (Picnic) contained scenes of eating and drinking, and the increased sales were probably due to the film itself rather than the subliminal advertising. This conclusion is based on the fact that there is very little evidence from over 200 studies that subliminal advertising is effective in changing behaviour (Pratkanis & Aronson, 1992).

Pratkanis, A. R., & Aronson, E. (1992). Age of propaganda: The everyday use and abuse of persuasion. New York: W.H. Freeman.

Stroop Effect

2010-02-02T08:18:00.000-08:00

The Stroop effect, in which the naming of the colours in which words are printed is slowed down by using colour words (e.g., the word YELLOW printed in red), seems to involve unavoidable and automatic processing of the colour words. However, Kahaneman and Henik (1979) found that the Stroop effect was much larger when the distracting information (i.e., the colour name) was in the same locations as the to-be-named colour rather than in an adjacent location. Thus, the processes producing the Stroop effect are not entirely unavoidable and so not completely automatic.

Sound Patterns

2010-02-02T08:17:00.000-08:00

(The Sound Shape of Language, Roman Jakobson and Linda R. Waugh)

"Sound Patterns in Language" (1925) was Edward Sapir's momentous contribution to the first issue of the first volume of the review 'Language', published by the newborn Linguistic Society of America. This first American pathfinder (1884-1939) in the theoretical insights into the sound shape of language said that "a speech sound is not merely an articulation or an acoustic image, but material for symbolic expression in an appropriate linguistic context"; and it was on "the relational gaps between sounds of a language" that Sapir put the chief emphasis. Similarly, the topological idea that in any analysis of structure "it is not things that matter but the relations between
them", an idea which found a manifold expression in contemporaneous sciences and arts, was a main guide for the exponents of the Prague Linguistic Circle, founded in 1926. They endeavored to derive the characteristics of phonemes from the interrelations of these units and in the "Project of Standardized Phonological Terminology" of 1930 they defined a 'phonological unit' as a term of an opposition. The concept of 'opposition' took on fundamental importance for the differentiation of cognitive meanings. The question of the relationship between the sense-discriminative units bacame the necessary requirement for any delineation of functional sound systems.

Sociolinguistics of Writing

2010-02-02T08:16:00.000-08:00

'For a long time, writing was a secret tool. The possession of writing meant distinction, domination, and controlled communication, in short, the means of an initiation. Historically writting was linked with the division of social classes and their struggles, and (in out country) with the attainment of democracy.' (Roland Barthes)

'Literate societies are characterized by a literate environment which promotes extensive and regular use of literacy in all communicative domains. In such societies, illiteracy is considered to be a stigma by both the literate and the nonliterate sections of the society.' (Chander Daswani)

Language is a social fact, which implies that it is a mental phenomenon. Its written form speaks to me the mind in its own way, shaping the language users' awareness of their language and hence its identity.

Signs of Words

2010-02-02T08:13:00.000-08:00

All words of necessary or common use were spoken before they were written; and while they are unfixed by any visible signs, must have been spoken with great diversity. -- Samuel Johnson

It would be necessary to search for the reason for dividing language into words - for in spite of the difficulty of dealing it, word is a unit that strikes the mind, something central in the mechanism of language. -- Ferdinand de Saussure

Theoretical words

Words are typical units of lexicology and lexicography. This seems obvious enough, but there has been a great deal of scholarly discussion about the status of the word in language structure. Some linguists avoid the term altogether giving preference to the morpheme as the smallest and basic grammatical unit. For, while in everyday speech we can live with expressions that have vague and multiple meanings, scientific terms should be unambiguous and, ideally, universally applicable. The word fails on both counts. 'Word' is a highly ambiguous term and hard to define in a way valid for all languages. Words are units at the boundary between morphology and syntax serving important functions as carriers of both semantic (Sampson 1979) and syntactic (Di Sciullo and Williams 1987) information and as such are subject to typological variation. In some languages words seem to be more clearly delimited and more stable than in others. The structural make-up of words depends on typological characteristics of languages.

(note: each verb and its forms are the same word, or different words?)

The remarks by Johnson and Saussure quoted at the beginning of this chapter point to the important fact that words are intuitively given units but hard to pinpoint. Once fixed by visible signs, they acquire a corporeal existence. It should be borne in mind that first and foremost words are lexical units or lemmata, that is, analytic units of the written language.

Signs of Segments

2010-02-02T08:04:00.000-08:00

'Each natural language has a finite number of phonemes (or letters in its alphabet) and each sentence is represented as a finite sequence of these phonemes (or letters).' Noam Chomsky, Syntactic Structures

Segments, more specifically phonemic segments, are, it is widely believed, what alphabetic letters encode. However, alphabetic writing has been cited as evidence both for the psychological reality of segments (Cohn 2001: 198) and for the view that segments are a mere project (Morais et. al. 1979). The argument cuts either way. How would it be possible to encode speech as a sequence of discrete graphical elements (letters) unless there were corresponding units in the mental representation of language?! (...)

Theoretical segments

Phonologists define segments as ensembles of distinctive features referring to manner and place of articulation. These features are the cornerstone of phonological theory. Their combinations yield segments called 'phones' when they are not viewed as elements of a particular language. (...)

The production of phonetic features in connected speech extends over a period of time, starting before a segment begins and coming to an end only after it has been terminated (Günther 1988: 15). Where, then is the segment? According to Pierce (1992:384), 'one common intuition about talking is that we proceed by emitting a sequence of discrete articulation, rather like the letters of an alphabet'. It is quite common to equate segments with the letters of the alphabet in this manner, as witnessed, for instance, in the quote by Noam Chomsky at the beginning of this chapter. However, over the past several decades, phonologists have moved away from the segment, since they were not able to discern it in speech signal. Inspection of phonetic reality (connected speech) has not revealed segments corresponding to discrete phonemes (corresponding in turn to discrete graphemes), because articulators - that is, the physical organs of speech production - work continuously, exhibiting, at any point, the influence of the preceding and following sound. There is hence broad agreement that 'it is impossible, in general, to disarticulate phonological representation into a string of non-overlapping units' (Prince 1992:386). This is a real problem, for how shall we interpret letters as overlapping units? The problem disappears if, for descriptive purposes, we accept a model of language where there is a phonemic level, at which discrete segments are lined up one after another, as in writing.

All attempts to prove that speech actually works on the basis of principles determining the sequential organization of discrete segments have failed. At the same time, Chomsky's above-quoted statement that, on abstract level, speech is represented in terms of finite sequences of segments in indisputable. As a matter of fact, description of this sort have been highly successful. But a good description of an object need not be isomorphic with it. (...) In like manner we must not confuse a segmental description of speech with the speech itself. In a sense, alphabetic orthographies can be understood as descriptions of their respective languages, but in any event the relationship between sequences of alphabetic letters and speech is never a one-to-one mapping relation. It is complex in both directions, and, as any description, hinges on a certain point of view highlighting some aspects at the expense of others.

Alphabetically written words can be read and can be pronounced, even words like chlororophenpyridamine. The pronounceability of alphabetic words rests on a process known as 'phonological recording', that is, the transformation of mental representations of sequences of letters into mental representations of sequences of sounds. A great deal of reading research deals with the question of whether and to what extent phonological recording is necessary for reading alphabetic texts, a problem to which we will return in chapter 9. For present purposes suffice it to note the obvious fact that alphabetic texts can be given a phonetic interpretation, they can be read aloud. While this is true of all writing, more or less, it is widely assumed that in alphabetic writing this rests on the fact that each letter represents a sound. The question then is, what sound?

note: There are some words in our own language we don't know for sure how to pronouce or there are more than one acceptable pronounciation.

Phonemes

As pointed out above, the prime candidate, the phonetic segment or phone, has proven to be elusive. Phonologists have recourse to a more abstract unit, the phoneme defined as a phone which fulfils a meaning-differentiating function in a given language. Although there are problems with the phoneme, too, many phonologists continue to use this concept, telling us, for example, that on average languages have 22.8 consonants phonemes and 8.7 vowel phonemes. Maddieson (1984) reports these figures on the basis of studying 317 languages. While he found that they differed on a large scale in their sound inventories, distinguishing between a poor 6 and a luxurious 95 consonants and between an equally disparate 3 and 46 vowels, this is clearly an order of magnitude altogether different from that of words, morphemes and syllables, however counted. In this regard, Cicero's (106-43 BCE) Latin was a plain-vanilla language. With 28 phonemes it is pretty close to the average. What this means is that in sound pattern of first-century Latin we find 28 important contrasts that are systematically used to differentiate meaning. A contrast is not the same as a unit, although this distinction is often ignored. Consider, for example, the following definition.

Segmental phoneme: a consonant and vowel sound of a language that functions as a recognizable, discrete unit in. To have phonemic value, a difference in sound must function as a distinguishing element marking a difference in meaning or identity. (Ives, Bursuk and Ives 1979:253)

The difference between a unit and a contrast is often glossed over like this because our inability to pin down the segment can thus be concealed. It is, however, possible to give every contrast a name, say a letter, which is then used to mark it. This kind of relationship between phonological distinctions and letters has often been interpreted as meaning that 'the purpose of alphabetic orthographies is to represent and convey phonologic structures in a graphic form' (Frost 1992: 255). Who, if anybody, stipulated this purpose is unknown. If orthographies have a purpose it is to encode and retrieve linguistic meaning in a graphic form. To represent and convey phonological structure is at best a means to that end, which is of no interest to anyone except linguistic. Instead of assuming a purpose at all it seems more prudent to consider an alphabetic orthography as a possible interpretation or description of the phonological structures of a language, and not usually an ideal one for that matter, if by ideal we mean being parsimonious and as simple as possible.

(...)

Written segments

(...) In Johnson's day, a letter was a thing with three attributes, a name (nomen), a graphical form (figura) and a power (potestas), that is, its pronunciation. Form and name relatively unproblematic, but the power was 'vague and unsettled'.

Uncertainty and polyvalence

This uncertainty has three aspects. One is that, even assuming that each letter of the Latin alphabet was interpreted as a phoneme, these interpretations were clear only as contrasts, that is within the system of Latin phonology as reflection in spelling. Secondly, some uncertainty is bound to arise whenever the letters whose phonetic correspondences are defined with respect to the relevant contrast of one language are applied to another where at least some of the contrasts are different. There is no complete congruence. Finally, there is the uncertainty of which contrast are relevant in the hitherto unwritten language and how they should be marked.

(...)

Historical change

Over time, the gap between spelling and pronunciation is bound to widen in alphabetic orthographies, as spoken forms change and written forms are retained. Many of the so-called 'silent' letters in French can be explained in this way. Catach (1978:65) states that 12.83 per cent of letters are mute letters in French, that is, letters that have no phonetic interpretation whatever. Many of them once had phonetic counterparts that, by regular processes of sound change, have been effaced. (...)

Another historical factor that undermined simplicity and cross-linguistic uniformity in sound-letter correspondences of the Latin alphabet has to do with the gradual reversal of the relationship between speech and writing. 'How shall I write this word?' used to be the initial question where the application of the Latin alphabet to an unwritten language was at issue. As time went by, it was superseded by the question 'How shall I pronounce this word?' (...) writing had become an agent of linguist change, transcending its role as a means of expression. The image become the model.

Some linguists consider that this is an inevitable consequence of writing, as, for example, the title of Kenneth Pike's 1947 book suggests: Phonemics: A Technique for Reducing Languages to Writing. Phonemes are here seen in direct correlation with alphabetic writing, which, from Pike's point of view, is a reduction, an abstraction, rather than a neutral and faithful representation. A letter is a stabilizer, something like a catalyst, which introduces shape where in phonic reality is flux. It is worth nothing that this is a problem not just of description, but of standardization and the power of a fixed norm. Writing by means of letters that supposedly represent sounds fosters an awareness of the necessity to settle on a variety embodying the canonical form of the language in question.

Suprasegmental

Sound features such as stress and pitch are essential parts of utterances, but the Latin alphabet provides no means of encoding them. These features are called 'suprasegmentals' because they do not occur before or after, but together with other vowels and sonorants. They relate not to segments but to syllables and sometimes larger units.

Conclusion

The Latin alphabet is the most widely used script of all time. Its simplicity and elegance as the writing system of the Latin language suggests universal applicability on the basis of the common principle of segmentation. More than any other script it is associated with the idea of the sound segment.

Signaries and Statistics

2010-02-02T08:00:00.000-08:00

It is obvious that the more complexity of the possible syllables of a language interacts with their numbers. A language such as Fijian that permits only open syllables is bound to have a fewer syllables than one that permits syllables with complex initial and final margins of the type of English strength. Also it would appear that a language whose basic lexical stratum is monosyllabic needs more syllable types than one that has a basic stratum of polysyllabic lexemes. A writing system that targets the syllable as the key function unit thus means different things for different languages. (...)

------------------------

The Chinese Script Reform Committee says that there are more than 1,200 syllables in Mandarin, this total number of possible syllables is much smaller than that in English. (...)

------------------------

Syllable Structure: The Limits of Variation
By San Duanmu

The total number of possible syllables, therefore, should be the number of possible onsets times the number of VX rhymes times the number of choices for the extra final C.

Possible monosyllables in English. Total = Onsets x VX x C = 59 x 298 x 10 = 170,510

------------------------

Clearly, this is an order of magnitude that makes syllabaries unmanageable. In practice there are no, and never have been any, complete syllabaries in the above sense, which confirms the more general truth that no writing system encodes every distinction relevant in its language. Various strategies were developed for syllabic writing to get by with signaries such smaller that the number of speech syllables. An inevitable consequence of this is a certain degree of syntagmatic complexity in combining graphic symbols unambiguously to denote speech syllables.

Where syllabic writing evolved, the number of symbols was gradually reduced. (...)

Much has been made of the economic advantage of syllabic writing over word writing, which stems from the fact that the number of the speech syllables of a language is closed while that of words is open. Gelb (1963) in particular considered the economizing on the inventory of signs was the driving force in the development of writing. This is the cornerstone of his theory. To be sure, the structural unit of writing has an effect on the size of a writing system's signary.

Hearer's Apprehension

2010-02-02T07:57:00.002-08:00

(The Sound Shape of Language, Roman Jakobson, Linda R. Waugh)

Indisputably, the grammatical pattern of the sentence, the verbal context of the words at issue, and the situation which surrounds the given utterance prompt the hearer's apprehension of the actual sense of the words so that he doen't need to pick up all the constituents of the sound sequence.

Reading Saussure

2010-02-02T07:57:00.001-08:00

(Roy Harris)

'Reading Saussure' might perhaps be regarded as controversial title for a study of a book which Saussure never wrote. In one sense, we can no more read Saussure who was the founder of Saussurean linguistics than we can read the Socrates who was the founder of Socratic philosophy. (...) (It may come as something of a shock today to realize that none of Saussure's three courses at Geneva was attended by more than a handful of students.)

Are the key concepts of the 'Cours' to be viewed as deriving specifially from the work of Humboldt, or Paul, or Gabelentz, or Durkheim, or Whitney ...? Or were they, as Bloomfield brusquely claimed in his review of the book (Bloomfield 1923), just ideas which had been 'in the air' for a long time?

Reaction Time

2010-02-02T07:55:00.000-08:00

(...) assessing visual search performance only by reaction time (as is generally done) is limited, because speed of performance depends partially on the participants' willingness to accept erros.

(...)
As McElree and Carrasco (1999, p.1532) pointed out, "RT [reaction time] data are of limited value ... because RT can vary with either differences in discriminability, differences in processing speed, or unknown mixtures of the two effects."

(...)
As Wolfe (1998, p.56) pointed out:

"In the real world, distractors are very heterogeneous [diverse]. Stimuli exist in
many size scales in a single view. Items are probably defined by conjunctions of many
features. You don't get several hundred trials with the same targets and distractors... A truly satisfying model of visual search will need... to account for the range of real-world visual behaviour."

Information Rate

2010-02-02T07:52:00.000-08:00

(Lawrence Rabiner)

(...) Here we see the steps in the process laid out along a line corresponding to the basic information rate of the signal (or control) at various stages of the process. The discrete symbol information rate in the raw message text is rather low (about 50 bps [bits per second] corresponding to about 8 sounds per second, where each sound is one of about 50 distinct symbols). After the language code conversion, with the inclusion of prosody information, the information rate rises to about 200 bps. Somewhere in the next stage the representation of the information in the signal (or the control) becomes continuous with an equivalent rate of about 200 bps at the neuromuscular control level, and about 30,000-50,000 bps at the acoustic signal level.
(...)
The steps in the speech-perception mechanism can also be interpreted in terms of information rate in the signal or its control and follows the inverse pattern of the production process. Thus the continuous information rate at the basilar membrane is in the range of 30,000-50,000 bps, while at neural transduction stage it is about 2000 bps. The higher-level processing within the brain converts the neural signals to a discrete representation, which ultimately is decoded into a low-bit-rate message.

Hilary Putnam

2010-02-02T07:51:00.000-08:00

„In his John Locke lectures, Hilary Putnam argues „that certain human abilities – language speaking is the paradigm example – may not theoretically explicable in isolation“, apart from a full model of „human functional organization“, which „may well be unintelligible to humans when stated in any detail.“ The problem is that „we are not, realistically, going to get a detailed explanatory model of the natural kind „human being“ not because of „mere complexity“ but because „we are partially opaque to ourselves, in the sense of not having the ability to understand one another as we understand hydrogen atoms.“ This is a „constitutive fact“ about „human beings in the present period, though perhaps not in a few hundred years.“ (Chomsky 2000: 19).

Psycholinguistics of Writing

2010-02-02T07:47:00.000-08:00

(Writing Systems: An introduction to their linguistic analysis, Florian Coulmas)

'When he was reading, his eye glided over the page, and his heart searched out the sense, but his voice and tongue were at rest. (Augustine)

'Writing requires deliberate analytic action on the part of the child. In speaking, he is hardly conscious of the sound he produces and quite unconscious of the mental operation he performs. In writing, he must take cognizance of the sound structure of each word, dissect it, and reproduce it in alphabetic symbols, which he must have studied and memorized before.' (Lev S. Vygotsky)

(...) the introduction to writing implies a cognitive reorientation and restructuring of symbolic behaviour. Names of objects are conceptually dissociated from their denotata, as signs of physical objects are reinterpreted as signs of linguistic objects, names. In a second step, signs of names are recognized as potentially meaningless signs of bits of sounds, which are then broken down into smaller components.

Reading

The bulk of all reading research is concerned with writing systems that make use of the alphabetic notation. (...) It should be kept in mind, however, that this focus on the alphabet has implications for the questions that are asked, how they are pursued, and eventually for theory formation.
(...) linguists and philologists have described and classified writing systems variously as logographic, ideographic, morphosyllabographic, syllabic, phonemic and so on. These classifications are one thing; but how writing systems work in terms of actual perception, processing and production is another. Psycholinguistic research into reading can shed new light on classifications derived from structural descriptions, and lead to a reassessment of how meaningful they are.

Word superiority

In antiquity, texts were commonly redacted in 'scriptura continua', without word
boundaries (Saenger 1991). (...) words, unlike speech sounds, are meaningful, and this is what reading is all about. We read not to intone, but to understand.

The minimal coding unit of alphabetic writing systems is smaller than the word, but modern alphabetic texts consists of words divided by spaces, reflecting the intuitive insight that word separation facilitates reading. The reader's general task is to 'search out the sense' that is linguistically encoded.' (...) Letters are recognized more quickly and more accurately when presented within words (e.g. input) than in isolation or within pseudowords (e.g. inpat). This finding leads to the concept of a lexicon or mental dictionary against which the visual input is matched. In fluent reading, a visual input is linked to a lexical entry that contains morphological and semantic information such as the part of speech of the word and its meaning.

Stroop (1935) discovered that naming the colour of the ink in which a word is written is delayed when that word is the name of a different colour. Stroop, J. Ridley. 1935. Studies of interference in serial verbal reactions. Journal of Experimental Psychology 18, 643-62.

Reading acquisition

The conflicting views about the role of phonological recording in fluent reading are mirrored in a long-standing controversy that pervades reading teaching methods. On one hand, the phonics and decoding method views reading as a process that converts written forms of language to speech forms and then to meaning. A teaching method, consequently, should emphasize phonological knowledge. As one leading proponent of the phonic/decoding approach puts it, 'phonological skills are not merely concomitants or by-products of reading ability; they are true antecedents that may account for up to 60 per cent of the variance in children's reading ability' (Mann 1991: 130). On the other hand, the whole-word method sees reading as a form of communication that consists of the reception of information through the written form, the recovery of meaning being the essential purpose. 'Since it is the case that learning to recognize whole words is necessary to be a fluent reader, therefore, the learning of whole words right from the start may be easier and more effective' (Steinberg, Nagata and Aline 2001:97).

(...) Phonics and decoding advocates do recognize that learning sight words may be functional in reading acquisition, and, by the same token, whole-word advocates do not deny that the teaching of sound values of letters can serve a useful purpose in reading instructions. Fluent readers rely on different strategies for word recognition: matching sight words with templates stored in memory; predicting words from context; applying grapheme-phoneme correspondences to reconstruct phonological words; guessing unknown words by analogy to others already known. Moreover, from research that focussed on the reading of longer words it follows that lexical access is not a process involving only letters and words. Subword units, especially morphemes, and also functional (Taft 1987).

Reading methods

It seems that so far reading research has produced more questions than answers. This impression is partially correct, and there are two main reasons for it. One is that the full complexity of what happens between the stimulus of a piece of text hitting the retina and its meaning being interpreted in the brain is only gradually becoming apparent. The other has to do with the enormous difficulties of devising experiments from which conclusions can be drawn about this process. For on conclusions we have to rely, because direct observations is impossible.

Cognitive consequences of writing

Vygotsky was one of the first psychologists to take an active interest in the cognitive consequences of writing. Working in the 1930s, he was intrigued by the question of how awareness of the properties of speech is affected by writing. In the meantime, numerous studies ranging from the flow of ideas and the level of discourse planning (Chafe 1987) to that of speech-sound segmentation abilities (Morais 1987) have lent support to the notion that people's knowledge about language and their actual language use are influenced by literacy.

Phonetics and Phonemics

2010-02-02T07:44:00.000-08:00

(Language and Symbolic Systems, Yuen Ren Chao)

Phonetics may be compared to the lines of longitude and latitude drawn on the globe and phonemics to the mapping of actual continents and oceans and countries. The precise way in which the divisions are made is to some extent arbitrary. During the French Revolution, it was attempted, though without success, to change the quadrant of 90° into 100° decimal degrees. But certain features, such as the North and South Poles and the Equator, are a part the nature of things. Similarly, stops and continuants, voice and voicelessness are natural variables found in all human speech. In phonetics one tries to anticipate, after a broad survey of the accessible languages of the world, all the necessary distinctions and set up standard points (such as the cardinal vowels and the divisions along the roof of the mouth) and then assign the actual sounds of any language under study to the nearest standard points, with the appropriate IPA symbols, so as to have an accurate representation of the sound of that language.

(...) There is therefore really no conflict between the two points of view about a phoneme being a group and being a set of features. They are two sides of the same coin. In fact Bernard Russell long before the theory of phonemes had a theory of equivalence between the property of a class and class membership. To paraphrase his "principle of abstraction", we might say that humanity (in the abstract) is humanity (mankind). Applied to phonemics, we might say that the common property of a number of different sounds which makes them members of one phoneme consists in the fact that they belong to this class.

This evident circularity in characterizing the property of a phoneme by its members is unavoidable because if you stipulate that members of a phoneme must be phonetically similar, a condition often included in the definition of a phoneme, then you run into cases where what to foreigners seem very different sounds belong to the same phoneme and the differences are hardly noticeable to the native speaker. The solution to this problem, as to all solutions in science, is to make your circle of circularity as big as possible. One important step in carrying this out is to look for cases of what is known as 'complementary distributions'. If a dorsal stop occurs always with the palatal articulation when followed by a front vowel (as in key) and always by a velar articulation when followed by a back vowel (as in call) but never the other way round, there is a case of complementary distribution.

But complementary distribution alone is not sufficient to determine what sounds go together to be members of one phoneme. There must also be overall symmetry in the organization of sounds into phonemes. For example, besides the complementary distribution of the palatal consonant in key and the velar consonant in call, there is also a parallel difference in quality in he and fall. Likewise, we have parallel differences in the g of geese and gall. Thus, we arrive at a neat and symmetrical system of groupings. Similarly, not only is k aspirated when initial and stressed and unaspirated when following an s, but the same is true of t in term and steam and of p in peak and speak. On the other hand, no one would seriously make one phoneme out of two sounds [h] and [n] in English simply because [h] always occurs as a syllabic initial and [n] always as a syllabic ending. Not only are the two sounds extremely dissimilar phonetically, but there is no other parallel case of complementary distribution in the sounds of English.

To summarize, then, a phoneme can be defined as one of an exhaustive list of systematized classes of phonetically related sounds in a language, such that every form in the language can be given as a (usually serially ordered) set of one or more of these classes. As definitions go in matters concerning human behaviour, this definitions is no more than a summary of usage and procedure among linguists and the definition does not even guarantee that its applications will always result in one unique system for any give language.

Perception and Memory

2010-02-02T07:43:00.000-08:00

There is a further assumption of the levels-of-processing approach that deserves brief discussion. According to Craik and Lockhart (1972), memory traces can be regarded as records of analyses carried out during perception. As a result, brain areas involved in perception and storage of information should be reactivated when memory is tested. Brain-imaging research supporting that position is discussed by Nyberg (2002). For example, consider a study by Nyberg et al. (2003). Participants initially learned visually presented words paired with sounds. After that, they performed a recognition memory test on visually presented words paired with sounds. After that, they performed a recognition memory test on visually presented words, but were not required to remember any auditory information. In spite of that, there was increased brain activity in auditory regions of the temporal lobes. As Nyberg (2002, p.346) pointed out, this finding provides "storage evidence that perceptual information is part of memory traces".

Mixed systems

2010-02-02T07:35:00.000-08:00

(Writing Systems: An introduction to their linguistic analysis, Florian Coulmas)

Egyptian writing

Until its decipherment by the French philologist Jean François Champollion (1790-1832) in the early 1820s, the nature of Egyptian writing system eluded European scholars. The impressive naturalism of the pictographs as well as a number of misconceptions handed down from antiquity led them to believe that the script was symbolic and that each one of the so-called 'holy characters' - which is what 'hieroglyph' means - was to be interpreted as a word, if not an entire sentence. As it turned out, only few Egyptians signs are logograms. Champollion arrived at the conclusion that Egyptian hieroglyphs were a phonetic rather than symbolic script when he analysed the Rosetta Stone, a stele dating from 196 BCE, which is inscribed in two languages, Greek and Egyptian, and three scripts, the Greek alphabet, and the hieroglyphic and demotic varieties of Egyptian. Counting Egyptian signs and Greek words, he found that there were 1,419 signs for 486 words. Since the 1,419 were made up of only 66 different signs, he correctly concluded that the script was phonetic, at least in part.

After Champollion's path-breaking discovery, the detail of the system were worked out. It is a mixed system consisting of phonograms of various sorts and signs that are interpreted for meaning rather than sound. All in all some 700 signs were used to write the language during its classical period in the second millennium BCE, although the total number increased significantly in later periods. However, more than 400 signs were rarely needed at any one time. In addition to the pictorial hieroglyphs that are always used to monumental inscriptions, the Egyptians developed two more cursive script forms for manuscript writing, which allowed for greater speed and efficiency: the semicursive 'hieratic' and the cursive 'demotic', both so called by the Greeks. The iconic properties of the hieroglyphs were lost in these cursive scripts, but structurally they remained fairly close to the hieroglyphic model.

One important difference between the hieroglyphic script and the two cursive scripts has to do with the grouping and combination of signs. While hieratic and demotic writing is always linear from right to left, there is often no such clear correspondence between the flow of speech and the linearity of the hieroglyphic script. The arrangement of hieroglyphs makes more liberal use of the two dimensions of the writing surface, often closely integrating the written text with graphic images relating to its contents. Both the orientation of the hieroglyphs and their spatial arrangement in columns and horizontal lines is more varied and flexible than hieratic and demotic writing. Hieroglyphs are sometimes switched in their order for reasons of better spacing (Davies 1987: 13).

A defining feature of the Egyptian writing system is that it provides no overt cues for the interpretation of vowels (Schenkel 1984). Egyptian belongs to the Hamito-Semitic family of languages, which is characterized by triconsonantal word-roots. Other writing systems for Semitic languages also focus on consonants but like the North Semitic cuneiform systems and the consonant alphabets of the Phoenician family of scripts, the Egyptian script has no auxiliary means of vowel indication. On the basis of consonantal frames readers supply the contextual appropriate vowels to interpret the full body of the word.

To this end they make use of the three classes of signs that, as all available evidence suggests, were present in the earliest stages of the Egyptian script around 3000 BCE (Fischer 1989) and used continuously until the end of its long history in the tenth century CE. There three classes of signs are logograms, phonograms and determinatives.

Japanese writing

When the art of writing spreads across language boundaries from literate to non-literate cultures, it is common that the written language is adopted along with the writing system. This is what the Akkadians did when they mastered Sumerian cuneiform, and this is what the Japanese did when they learned Chinese characters. Language and scripts cannot be separated easily at first. To both the Akkadians and the Japanese writing for a long time meant writing a foreign language, Sumerian in one case, Chinese in the other. Once they began to write their own language, they adapted the existing system rather than creating a new one. Yet, the resulting new system departs from the structural make-up of its model in fundamental ways. Although there are many differences in detail between the adaptation of Sumerian cuneiform and Akkadian and that of Chinese characters for Japanese, some basic parallels are also in evidence. Since, with the exception of a handful of original creations (cf. chapter 10), all writing systems past and present are adaptations, it is of general interest to see what these parallels are.

The scribes who first used cuneiform signs to write Akkadian, and Chinese characters to write Japanese adjusted a fully developed writing system to a typological different language. Three mechanisms are involved in this process.

(1) Extant signs are reinterpreted. This is of some importance for the perspective on writing that informs this book. That is, the graphic sign precedes its linguistic interpretation. As we have seen in the previous section, the Akkadians took Sumerian logograms and assigned them an additional lexical interpretation, typically the Akkadian translation equivalent of the Sumerian word. The Japanese did exactly the same, providing Chinese characters with Japanese interpretations, while holding on the Chinese interpretation, too. Notice that assigning an extant sign an extra interpretation is conceptually quite different from using a sign to represent a word. In the latter case the underlying question is, 'How do I write this word?' But this is not primarily what the Akkadian and the Japanese scribes asked themselves. Rather, their point of departure was the sign, which they changed by giving it an additional interpretation it did not have in the original system. As a result, Chinese characters have a Chinese reading, called Sino-Japanese, and a Japanese reading in Japanese, just as cuneiform signs have a Sumerian reading and an Akkadian reading in Akkadian.

(2) Sings are used for their phonetic interpretation only. Conditions are a bit different here, because the Sumerians were using cuneiform signs both as logograms and as syllabograms before the Akkadians adopted the system, while purely phonetic usage of Chinese characters was more limited in Chinese. Yet, like the Akkadians the Japanese used the adopted signs for their syllabic values only, disregarding their meaning in Chinese. But while the Akkadians left the form of Sumerian syllabograms unchanged, although they made some adjustments for their sound values, the Japanese gradually changed the graphic shape of the Chinese characters they used for their syllabic values only, eventually developing a set of signs immediately recognizable as syllabograms. These are the two kana syllabaries discussed in chapter 4. Once again it is clear that extant signs were reinterpreted and subsequently graphically modified so as to mark them off as a functionally distinct set of signs not to be confused with Chinese characters.

(3) New signs modelled on those of the adopted systems are created. There are some cuneiforms logograms in Akkadian that do not exist in Sumerian. They were created by the Akkadians and are sometimes called 'artificial' Sumerograms (Krebernik and Nissen 1994). In the same manner, the Japanese created some new characters, applying the principles of graphic composition of Chinese characters. In contradistinction to the Chinese characters that were reinterpreted, these 'kokuji' or 'native characters' have no Chinese reading (although there are some exceptions where a pseudo-Chinese reading was invented). Many 'kokuji' were created for Japanese words that lack obvious Chinese translation equivalents, thus bearing witness to the writer's need to adapt the system to the Japanese language. Obviously, the same words can be written syllabically, but in many regards logographic writing was more appealing to the Japanese scribes, as it was to their Akkadian colleagues. Some examples of 'kokuji' are given in table 9.5. Notice that readings consists of two to four syllables, as a typical of Japanese words, whereas Chinese characters are consistently interpreted as one syllable in Chinese.

English writing

In contradistinction to Japanese, the English writing system has a defining unit, or so it would seem. On the face of it there is just one class of signs, alphabetic letters. But this is deceiving, because an analysis of English writing gets nowhere if it starts out with the assumption of individual letters having uniform canonical interpretations. In fact, the English writing system has long been recognized as a mixed system, although this has only rarely been pointed out explicitly (Stubbs 1996). It is mixed in the sense that there are units of different kinds and that the conventions for their interpretation are diverse. Albrow (1972), for example, identifies three sets of rules that operate in the English orthography: basic, romance and exotic. Chomsky and Halle's (1968) important work on the sound pattern of English is often quoted as hailing English writing as a near-optimal morphophonemic orthography, but building on this research Klima (1972) argued that there are at least six ways of analysing the English writing system, ranging in abstractness from phonetic to morphophonemic. Thanks to the multifaceted history of English writing (Scragg 1974) several different sets of rules combine, and sometimes compete with each other. And when all rules are exhausted, a considerable area of unpredictable spelling remains. (...)

Since the English writing system makes use of the Latin alphabet, the most common approach to its analysis is to figure out how these letters are to be interpreted. Attempts at compiling a comprehensive list of correspondence between sound and spelling of English have yielded various results. According ti Dewey (1971), the typical vowel can be spelt around twenty different ways. Later researchers such as Nyikos (1988) have found that the forty-odd phoneme of English correspond to 1,120 different graphemes. That is, on average twenty-eight different different letters and letter combinations are given the same phonetic interpretation. It is often assumed that letter-to-sound correspondences and sound-to-letter correspondences are mirror-image processes.

As a general rule, spelling conventions, once established, are more resistant to change than speech, which is another way of saying that written words tend to have an autonomous existence and phonetic interpretations are adjusted. Since sound changes, though regular, are contextual, not all words in which a certain sound occurs are affected in the same way. Individual letters are, therefore, bound to become more polyvalent in the course of time. Graphic autonomy is, indeed, quite important in English spelling where morpheme constancy is often given priority over phoneme constancy.

Formal Requirements for the Average Uncertainty

2010-02-02T07:33:00.000-08:00

(An Introduction to Information Theory, Fazlollah M. Reza)

Shannon's approach, as well as several other authos', in suggesting a suitable H function has been to some extent directed toward an axiomatic description of such functions. The desired H function should habe the following basic properties:

1. Continuity.
That is, if the probabilities of the occurrence of events Ek are slightly changed, the measure of uncertainty associated with the system should vary accordingly in a continuous manner. (...) a slight change in the probability of the occurrence of an event should not provide us with a significantly large amount of information.

2. Symmetry.
The H function must be functionally symmetric in every pk. Indeed, the measure of uncertainty associated with a complete probability set [Ek, Ek'] must be exactly the same as the measure associated with the set [Ek',Ek]. Our measure must be invariant with respect to the order of the events.

3. Extremal Property.
When all the events are equally likely, the average uncertainty must have its largest value. In this case, it is most uncertain which event is going to occur. Conversely, once we know which specific event among a number of n equally likely events has occured, we have acquired the largest average amount of information relevant to the occurance of events of a universe consisting of n complete events.

4. Additivity.
Suppose that we habe obtained a suitable measure of the average uncertainty H(p1,p2,...,pn) associated with a complete set of events. Now, let us assume that the event En is divided into disjoint subsets such that
(...)

Complying with properties 1 to 4 given above, or with similar requirements, one should be able to derive a functional form for the desired uncertainty function. Such treatments have appeared in the work of Feinsten, Khinchin, Shannon, Schtzenberger, and others. (...)
1. Fadiev assumes properties 1, 2, and 4, subsequent to several lemmas, proves that H must be of the form suggested in Eq.(3-11) expectfor a multiplicative constant.
H(X) = - sum_{k=1}^{n} p_k log p_k
2. Khinchin assumes properties 1, 3, and 4 and the fact that adding a null set to a complete set of events should not change its entropy, and he derives the form of Eq.(3-11) up to a positive constant multiplier.
3. Schutzenberger aims for a more general axiomatic search for a measure of information associated with a complete set of events. He shows that functions other than the Shannon-Wiener entropy of Eq.(3-11) may also be employed. An example of such a function is given in the work of R. A. Fischer. It should be pointed out, however, that the Shannon-Wiener suggested form is certainly the simplest of all such forms. The present richness and depth of the literature of information theory are to a great extent due to the simplicity of the form of Eq.(3-11).

A Quantitative Measure of Information

2010-02-02T07:29:00.000-08:00

(An Introduction to Information Theory, Fazlollah M. Reza)

In our study we deal with ideal mathematical models of communication. We confine ourselves to models that are statistically defined. That is, the most significant feature of our model is its unpredictability. The source, for instance, transmits at random any one of a set of prespecified messages. We have no specific knowledge as to which message will be transmitted next. But we know the propbability of transmitting each message directly, or something to that effect. If the behavior of the model were predictable (deterministic), then recouse to measuring an amount of information would hardly be necessary.

When the model is statistically defined, while we have no concrete assurance of its detailed performance, we are able to describe, in a sense, its "over-all" or "avarage" performance in the light of its statistical description. In short, our search for an amount of information is virtually a search for statistical parameter associated with a probability scheme. The parameter should indicate a relative measure of uncertainty relevant to the occurrence of each particular message in the message ensemble. We shall illustrate how one goes about defining the amount of information by a well-known rudimentary example. Suppose that you are faced with the selection of equipment from a catalog which indicates n distinct models:

[x1, x2, ..., xn]

The desired amount of information I (xt) associated with the selection of a particular model xk must be a function of the probability of choosing xk:

I(xk) = f(P{xk})

If, for simplicity, we assume that each one of these models is selected with an equal probability, then the desired amount of information is only a function of n.

I1(xk) = f(1/n) (1-2a)

Next assume that each piece of equipment listed in the catalog can be ordered in one of m distinct colors. If for simplicity we assume that the selection of colors is also equiprobable, then the amount of information associated with the selection of a color cj among all equiprobable colors [c1, c2, ..., cm] is

I2(cj) = f(P{cj}) = f(1/m) (1-2b)

where the function f(x) must be the same unknown function used in Eq. (l-2a).

Finally, assume that the selection is done in two ways:
1. Select the equipment and then select the color, the two selections being independent of each other.
2. Select the equipment and its color at the same time as one selection from mn possible equiprobable choices.

The search for the function f(z) is based on the intuitive choice which requires the equality of the amount of infomation associated with the selection of the model xk with color cj in both schemes (l-2c) and (1-2d).

I(xk and cj) = I1(xk) + I2(cj) = f(1/n) + f(1/m) (1-2c)

I(xk and cj) = f(1/mn) (1-2d)

Thus
f(1/n) + f(1/m) = f(1/mn)

This functional equation has several solutions, the most important of which, for our purpose, is
f(x) = -log x

To give a numerical example, len n = 18 and m = 8.
I1(xk) = log 18
I2(cj) = log 8
I(xk and cj) = I1(xk) + I2(cj)
I(xk and cj) = log 18 + log 8 = log 144

Thus, when a statistical experiment has n equiprobable outcomes, the average amount of information associated with an outcome is log n. The logarithmic information measure has the desirable property of additivity for independent statistical experiments. These ideas will be elaborated upon in Chap 3.

Meaning

2010-02-02T07:28:00.000-08:00

So far we have treated meanings of linguistic forms as parts of the external world of which they are symbols. The word 'dog' means the animal dog. The word is said to 'refer to', or 'denote', the thing and the thing is the 'referent' or 'denotatum'. But much of the meaning of language has to do with the attitude of the speaker toward the referent, toward the person spoken to, and toward his own act of speaking. This makes meaning in language a much more complicated matter than just symbols for things and of course much more interesting (...)

In a monolingual dictionary, every word is defined in terms of other words in the language. But if the user does not know the meaning of any of the words, the whole thing is defining in a circle, as we noted in the case of grammatical forms.

(Language and Symbolic Systems, Yuen Ren Chao)

The Strategy of Phonemics

2010-02-02T07:26:00.000-08:00

(Morris Halle)

In a recent paper B. Mandelbrot has showm that it is impossible to account for the high resistance to noise of speech on the basis of a continuous view. If linguistic messages (utterances) be thought of as continuous, the correction of erros in the reception cannot begin until the entire message is received, which would make correction well nigh impossible, certainly infinitely more difficult than it actually is. On the other hand, if a discrete view be adopted, correction of errors can begin upon receipt of each discrete unit (quantum), since the discrete units in the language are just a small fraction of all possible things that the ear can receive.

Mandelbrot investigated in detail the consequences of the discrete character of language only on one level, that of words. The necessity for discrete units on the other levels is implicit in his argument. The words themselves are thus viewed as being composed of discrete components, usually known as morphemes, which in turn consist of other discrete units, the phonemes.

Uncertainty-based Information

2010-02-02T07:25:00.001-08:00

(George J. Klir, Mark J. Wierman)

(...) it was the common belief of many scientist that the level of complexity we can handle is basically a matter of the level of the computational power at our disposal. Later, in early 1960s, this belief was replaced with a more realistic outlook. We began to undestand that there are definite limits in dealing with complexity, which neither our human capabilities nor any computer technology can overcome. One such limit was determined by Hans Bremermann [1962] by simple considerations based on quantum theory. The limit is expressed by the proposition: "No data processing system, whether artificial or living, can process more than 2x10^47 bits per second per gram of its mass." (...)

Using the limit of information procssing obtained for one gram of mass and one second of processing time, Bremermann then calculates the total number of bits processed by a hypothetical computer the size of the Earth within a time period equal to the estimated age of the Earth. Since the mass and age of the Earth are estimated to be less than 6x10^27 grams and 10^10 years, respectively, and each year contains approximately 3.14 x 10^7 seconds, this imaginary computer would not be able to process more than 2.56 x 10^92 bits or, when roimding up to the nearest power of ten, 10^93 bits. The last number — 10^93 — is usually referred to as Bremermann's limit
and problems that require processing more than 10^93 bits of information
are called 'transcomputational problems'.

Phonology and Language Use

2010-02-02T05:41:00.000-08:00

(Joan Bybee)

1. Language Use as Part of Linguistic Theory

1.1 Substance and Usage in Phonology

This book introduces into the traditional study of phonology the notion that language use plays a role in shaping the form and content of sound systems. In particular, the frequency with which individual words or sequences of words are used and the frequency with which certain patterns recur in a language affects the nature of mental representation and in some cases the actual phonetic shape of words.

Structuralism provided linguists with a workshop of analytic tools for breaking down the continuous speech stream into units, and these units into features; structuralism postulated hierarchical relations among the units and assigned structures to different levels of grammar, organizing language and the people who study it into subfields – phonology, morphology, syntax, and semantics.

The present work proposes to demonstrate that the focus on structure needs to be supplemented with a perspective that includes more than just structure, a view that includes two other important aspects of the language phenomenon – the material content or substance of language, and language use. The substance of language refers to the two polar ends – phonetics and semantics – that language molds and structures, the two ends between which language forms the bridge. Language use includes not just the processing of language, but all the social and interactional uses to which language is put. For present purposes, in the context of phonology, the frequency with which certain words, phrases, or patterns are used will be shown to have an impact on phonological structure.

Units and levels do not submit to definitions that work for every case. We still do not have strict definitions of even the most basic units, such as segment, syllable, morpheme, and word.

While substance has found its way into phonology from both the phonetic and semantic end, use has been systematically excluded from structuralist theories altogether. (...) totally excluding factors of use from consideration ignores the potential relation between representation and use. It is certainly possible that the way language is used affects the way it is represented cognitively, and thus the way it
is structured.

1.3 The Creative Role of Repetition

Haiman (1994, 1998) discusses grammar as ritualized behavior and points to various properties of both ritual and grammar that are the result of repetition. It is useful here to distinguish between a ritual and a convention: though both represent repeated behavior, a ritual can be individual and idiosyncratic, but a convention is agreed upon socially and evokes a consistent response in other members of a society (Tomasello et al. 1993). What both concepts have in common is that their structure is shaped by repetition. The following is a summary of some aspects of language that are shaped by repetition.

(...)repetition leads to strength of representation Repetition also leads to reduction of form. Repetition also leads to the reduction of meaning. Finally, and perhaps most importantly, repetition leads to emancipation.

Haiman (1994) demonstrates that the development of ritual is a common process in the animal kingdom argues (Haiman 1998) that ritualization is the basis for the development of grammar.

1.4.1 Token Frequency

Token frequency has two distinct effects that are important for phonology and morphology. In one frequency effect, phonetic change often progresses more quickly in items with high token frequency. This effect is particularly noticeable in grammaticizing elements or phrases that undergo drastic reduction as they increase in frequency. (...) But the effect occurs on a more subtle level as well: regular sound change in many cases progresses more quickly in items of high toke frequency.

1.4.2 Type Frequency and Productivity

Another major effect of frequency and thus of usage is the effect of type frequency in determining productivity. Productivity is the extent to which a pattern is likely to apply to new forms (e.g., borrowed items or novel formations). It appears that the productivity of a pattern, expressed in a schema, is largely, though not entirely, determined by its type frequency: the more items encompassed by a schema, the stronger
it is, and the more available it is for application to new items. (...)

The importance of productivity of both phonological and morphological schemas to our understanding of cognitive representations of language cannot be overstated. Productivity provides us with evidence about the generalizations that speakers make, and it is important to stress that speakers’ generalizations are not always the same as those devised by linguists on the basis of distributional evidence. Distributional evidence often recreates past stages of a language and does not reveal the restructuring and reanalysis that the patterns might have undergone. Productivity can be used as a diagnostic to determine which patterns are fossilized, and which represent viable schemas accessible to speakers.

1.4.3 Frequency Effects in Other Theories

The proposal that frequency of use affects representation suggests a very different view of lexical storage and its interaction with other aspects of the grammar or phonology than that assumed in most current theories. Structuralist and generative theories assume that the lexicon is a static list, and that neither the rules nor the lexical forms of a language are changed at all by instances of use. Similarly, as Pierrehumbert (1999) points out, all versions of Optimality Theory (Hayes 1999, Prince and Smolensky 1993, 1997) posit a strict separation of lexicon and grammar that makes it impossible to describe any of the interactions of phonology with the lexicon that are attested in the literature, many of which have just been mentioned: for instance,
the fact that many phonological changes affect high-frequency items first, and the fact that the strength of phonotactic constraints is directly related to the number of items they apply to in the existing lexicon. Hammond (1999) identifies an effect of frequency in cases of the application of the Rhythm Rule and proposes that an Optimality Theory account of the facts can include item-specific constraints inserted into the constraint hierarchy. But such a proposal neither fits well with other properties of Optimality Theory nor does it provide an account for why words and phrases of different frequencies of use behave differently.

1.5 Phonology as Procedure, Structure as Emergent

If we conceptualize phonology as part of the procedure for producing and understanding language, the phonological properties of language should result from the fact that it is a highly practiced behavior asso ciated with the vocal tract of human beings.
(...)

Children learn phonological sequences as parts of words, never independently of words. Articulatory routines that are already mastered are called forth for the production of new words, leading to a tendency of children to expand their vocabulary by acquiring words that are phonologically similar to those they already know (Ferguson and Farwell 1975, Lindblom 1992). This tendency leads to the structuring of the phonological sequences across words and the limiting of the potentially immense phonetic inventory. Put another way, the repetition of gestures and sequences across words allows relations of identity and similarity to develop in stretches of speech, giving rise to segment, syllable, and foot-sized units.

(...)

Grammatical and phonological structure emerge from the facts of co-occurrence in language use.

1.7 Language as a Part of Human Behavior

A basic assumption of this book is that the cognitive and psychological processes and principles that govern language are not specific to language, but are in general the same as those that govern other aspects of human cognitive and social behavior.

2 A Usage-Based Model for Phonology and Morphology

2.1 Introduction

The goal of phonology as conceived by generative theory is to describe the following phenomena: (i) the relations among similar but physically distinct sounds that are nonetheless taken to be ‘the same’ in some sense (allophonic relations), (ii) the relations among variants of morphemes as they occur in different contexts, (iii) phonological units of various sizes – features, segments, syllables, feet, and so on, and (iv) language-specific and universal properties of these relations and units.

(...) an alternate model, which does not make the same assumptions as generative phonology, can accomplish these same goals as well as accommodate other facts about phonological structure that are left unexplained in generative theory.

2.2 The Rule/List Fallacy

(...) Structuralist frameworks placed great emphasis on the systematicity of language, and it was thought appropriate to reduce the enormous complexity of language by extracting regularities that could be captured in general statements (i.e., rules) thereby only representing truly idiosyncratic material in a list (i.e., the lexicon). In this view, the major goal of linguistic analysis is to determine which features of a unit are idiosyncratic, and which are predictable by rule, with the additional desideratum of having as many features predictable by rule as possible.

However, there is no particular reason to believe that human language users organize or process linguistic material in this way.

The point is not to deny the presence of regularities, but rather to say that predictable features need not be excluded from representation in individual items. The presence of a feature on a list does not exclude it from being predictable by rule.

2.3 Organized Storage

Linguistic items are not stored in a long, unstructured list. Rather, the regularities and similarities observable in linguistic items are used to structure storage.

2.7 Units of Storage

The phonological shape of all words and frequent phrases that a person uses are stored in memory along with information about their meaning and contexts of use, both linguistic and nonlinguistic. The storage is not a simple list, but entails a network of connections to related items that makes storage more efficient.

2.8 Phonological Units

The articulatory and acoustic stream that constitutes speech is continuous and does not yield to exhaustive segmentation.

3.3 A Cognitively Realistic Model of Phonological Representation

Similarly, Miller (1994), summarizing the results of several experiments on perceptual categorization of phonetic segments, concludes that phonetic categories are both context-dependent and multiply specified.

3.4 Linguistic Evidence for Detailed and Redundant Storage

Even if one accepts the basic insight of the phonemic principle – that objectively different sounds form relations of similarity with one another – that does not mean that we must also accept the postulates that mental representations are phonemic (...).

3.4.1 Lexical Variation

There seems to be widespread agreement that phonetic change is gradual and produces variation. There also seems to be agreement that sound change interacts with the lexicon and morphology, changing lexical items and creating alternations.

3.5 Usage-Based Categorization versus Phonemic Representation

Any theory that proposes that stored representations differ from actual tokens of use must wrestle with the problem of the nature of the stored representation.

3.7 A Model for Sound Change

A usage-based theory postulates that change is inherent in the nature of language. Grammars are not static entities, but constantly in the process of change resulting from the way that language is used. Thus, a model of language must include the mechanisms by which change occurs as an integral part of its architecture. For that reason it is relevant – in fact, necessary – to provide an account of sound change in a model of usage-based phonology.

3.8 Special Reduction of High-Frequency Words and Phrases

It is also interesting to consider how words and phrases undergo reduction in the more extreme cases of contraction and grammaticization. One might want to argue that such cases are not central to a model of phonological representation, but such an argument would be based on the false assumption that such cases are not common. I would argue that such cases are quite common, and that new instances are arising all the time. Not only does special reduction occur in greetings and discourse markers, but it is also always present during the grammaticization process. Phonological reduction creates grammatical markers that continue to be a locus of change due to high frequency.

Language Transcription

2010-02-02T05:39:00.000-08:00

(Language and Symbolic Systems, Yuen Ren Chao)

(...) Henry Sweet (1845 - 1912) used to demonstrate that if you transcribe a language accurately, a person who has never heard the laguage before but knows the values of the symbols should be able to read if off and and make it sound like the original. On the other hand, a Swiss lecturer who spoke English once came to New York and delivered a lecture from notes written in the IPA and the audience could not understand a word he said. The incident does not prove inadequancy of transcription in gereral, but rather the incompleteness - even phonemic inompleteness - of simple segmental elements with omission of stress, length, intonation, etc. If a phonemic transcription is used, the reader will of course have to know the 'rules of pronunciation', namely which phonetic value of the phonemes is to be used.

Some Languages are Harder than Others

2010-02-02T05:37:00.000-08:00

(in Language Myths, by Lars-Gunnar Andersson)

We can begin by considering the sound system of languages. It must surely be the case that the fewer vowels, the fewer consonants and the simpler syllabic structure a language has, the simpler the sound system is. Hawaiian has thirteen distinctive sounds ('phonemes' in linguistic terminology), of which eight are consonants and five are vowels. Since the language also has stric rules about the syllable structure (almost all syllables have to consist of one consonant and one vowel in that order), the total number of possible syllables in the language is only 162. Compare English, where the consonants can be grouped together both before and after the vowel as in screams and splints. Of all the languages of the world, Hawaiian has one of the simplest sound system. At all other end of the scale we find Khoisan languages (previously known as Bushman and Hottentot languages). According to a recently published description, !Xóõ (that is actually how it is spelt), a language spoken in Southern Botswana, has 156 phonemes, of which 78 are rather unusual sounds called clicks, 50 are ordinary consonants and 28 are vowels. Studies of other languages in the are have also arrived at phoneme counts of around 150. The sound system of these languages are extremely complex.

(...)
We speak of analytic languages with little or no inflection and derivation and synthetic languages with a larger degree of inflection and derivation. (...) In absolute terms one could say that analytic languages are easier than synthetic languages, and there are two arguments for this claim. First, children always learn a more analytic version of their native language first; inflectional and derivational suffixes are learned later on. Secondly, pidgin languages from around the world are typically analytic. By pidgin languages we mean contact languages that arise or develop spontaneously. Most pidgin languages are found in the old European colonies around the world. One such language is Fanagalo, which has been used as a contact language between whites, blacks and coloureds in southern Africa since the nineteeenth
century, not least in the mining industry and in domestic services. (...) The world's most famous pidgin language speaker is Tarzan. When he says 'Me Tarzan, you Jane', he uses a simplified version of English.

(...)
One could, of course, object that pidgin languages are not real languages because nobody has them as a mother tongue. On the other hand, pidgin languages sometimes become the mother tongue of a group of people. They are the called creole languages. During the process of creolization, different complications in the grammar (as well as in the lexicon) will arise, but for a number of generations there creole languages will remain relatively simple. There is then good reason to believe that analytic languages are easier than synthetic. A more general conclusion could be that it is actually possible to speak of easier and harder languages with regard to grammar.

Language Changes

2010-02-02T05:36:00.000-08:00

Whether or not a country has an official standard for its national languages set and maintained by an academy or in the form of dictionaries and prescriptive grammars, the actual language of its speakers keeps changing, not only from generation to generation,
but also during the lifetime of a single person. Language being a set of habits maintained chiefly through the interaction between members of a speech community, it will change if the frenquency of intercommunication is diminished. Therefore, instead of asking why does a language change, a more natural question is to ask why should a language remain as stable as it is? Habits change, things are forgotten, people drift apart, and it is remarkable that language does not change faster than it does.

(Language and Symbolic Systems, Yuen Ren Chao)

Nonalphabetic Writing Systems: Some Observations

2010-02-02T05:31:00.000-08:00

(Samule E. Martin)

Most major languages in the world today are written with symbol systems of symbols that are similar in spirit to our own alphabet. The native writter makes use of a relatively small inventory of symbols to represent phonological elements -- typically syllables or phonemes, but in some cases morphophonemic and even componential elements. For various reasons, every writing system is in some way defective even when first instituted, and the institutionalization itself prevents the continuous revision that would keep the conventions of writing as well matched to the continuously changing spolen language as they may have been in the beginning. As a result, literate people tend to create formal written languages characterized by older forms, phonological and grammatical, which persist with very little change for long periods of time, until some sort of revolution demands a new written language that will correspond more closely to the living tongue.

The language spoken by more people than any other, Chinese, is usually written with a nonalphabetic system, a relatively large stock of characters that primarily represent words or parts of words. At the time the writing system developed, the overwhelming majority of Chinese words were single morphemes and the morpheme itself was in virtually all instances a single "syllable" -- traditionally defined as an initial plus an ending, the latter consisting of a vocalic nucleus with a tone and an optional coda. Accordingly, the writing has variously been called logographic or morphemic or monosyllabic, and there is no need to worry undduly about the slight inaccuracies implied by any of these terms. (The term "logographic" was popularized by Bloomheld and Kennedy, who seem to have taken it from Du Ponceau [1838:110]: "I would not call the Chinese chatacters a syllabic, but a logographic system of writing". Du Ponceau also use the term "lexigraphic.")

(...)

What are these Chinese characters, and how many of them are there? The K'anghsi dictionary of 1716 lists 40,545 different characters; Morohashi's recent dictionary carries nearly fifty thousand. (...) If we were to take all the characters that have ever existed, it is said, the total number would reach eighty thousand [Du Ponceau
1838, p.7n; Alleton 1970, p.47]. (...)

How are the Chinese characters structured? By tradition the graphs are divided into six categories, according to their origins; but the sixth category ("derivate characters") is so small and controversial that it is bets ignored. The five other categories are as follows:
(1) 'Pictographs' are direct iconic representations, such as those that depict the sun, the moon, a tree, a mouth, a mountain, a well, a bow, a stream, a gate, a shell, etc. Most characters have become highly stylized with the passage of time so that the original picture is not always obvious at first glance.
(2) 'Simple ideographs' depicts a logical idea: three horizontal lines to represent the number three, a pointer above or below the line to signal the words for up and down, etc.
(3) 'Compound ideographs' represent an abstract idea by combining two simple graphs, as when MOON is put to the right of SUN to represent the word for 'bright'. Two TREES are put together to represent the word 'grove'; three are combined to represent a different word for 'forest'.
(4) 'Phonetic loans' borrow a graph to represent a different word with the same or some similar sound, as when the character depicting a dustpan was borrowed to write the similar-sounding qí 'its', or the character depicting a kind of wheat was borrowed to write lái 'come'.
(5) 'Phonetic compounds' contain an element that hints at the meaning, usually called the 'radical', and an element called 'phonetic' that hints at the sound. The radicals, a kind of semantic classification system, were reduced in number from 540 of the first-century dictionary Shuõ Wén to the 214 found in the eighteenth century K'anghsi dictionary.
(...)

Is is said that 90 percent of all characters are phonetic compounds, with about 5 percent simple pictographs or ideographs and the remaining 5 percent compound ideographs or phonetic loans [Alleton 1970, p.33]. But in the modern shapes, both those created in China and those created in Japan, the phonetic and/or the radical sometimes disapears in the process of simplification. And we would expect the proportion of simple pictographs and ideographs to rise slightly as the number is restricted to the more frequent characters.

The shapes of the characters reflect a two-thousand-year history of brushwork calligraphy, of which one of the most important features is equidimensionality. Each character occupies an imaginary space of identical size, so that the same element must be given a smaller shape as it enters into more complex characters. (...)

For reasons apparently unknown, texts in East Asia have traditionally been written in vertical lines arrayed from right to left. But modern nonfiction publications have made increasing use of horizontal left-to-right printing, since that makes it easier to incorporate the foreign words and symbols so necessary in scientific work. (...)

Although Chinese writing is essentially logographic or morphemic, the characters are also phonetically to transliterate foreign names; these, however, are often abbreviated to the first syllable, especially when well known and when cited with a title. (...) Although the characters are not being used in their original senses, an attempt is made to choose characters that have a pleasant or neutral connotation, unless referring to a political personality, when the writer is free to vent his feelings by picking characters with abusive meaning.
(...)

When Chinese characters reached Japan they were eagerly embraced by an inquisitive people who admired discipline, respected learning, and showed an extraordinary taste for variety. Unfortunately they also had tin ears. The phonetic simplicity of their native language, with well under a hundred original syllables, made it a tortuous task to assimilate the more elaborate syllables of classical Chinese. Because of the richer syllable structure, Korean versions of Chinese words sound like Cantonese without tones; Japanese versions have a flavor all their own. The characters were repeatedly imported on different occasions and by different contacts with mainland; as a result, the Japanese ended up with more than one "Chinese" reading for many of the characters.
Now, the Chinese itself a number of characters have more than one traditional reading, the result of ancient derivational process or of dialect mixture.
(...)

When the Japanese borrowed Chinese characters they made an innovation that was to have far-reaching consequences. The characters represented Chinese morphemes, which the Japanese dutifully borrowed as words and as bound elements; but the morphemes themselves represented 'meanings', and the Japanese already had ways to express many of these meanings in their own, vastly different, language. So the Japanese started to associate a given character not only with the Chinese morpheme or morphemes imported with it but also with native words that were translational equivalents, calling these the "Kun" (explanatory) reading as contrasted with the "On" (phonetic) readings. (...) Since the exact misture of the three scripts depends upon an individual's education, knowledge, and whim, it is not surprising to find widespread anarchy in orthographic practices up until the educational reforms that followed Japan's defeat in World War II; these reforms have attempted to bring some semblance of standardization to the orthography, as well as a general simplification. Miller [1967, p. 133] tells us, "In 1927 the major Tokyo newspaper kept in stock printing type for between 7,500 and
8,000 different Chinese characters, and it was estimated at the time that an 'educated reader' would be 'familiar' with about 5,000 of these."
(...)

The postwar reforms aimed at reducing the number of characters and limiting the sanctioned readings to those deemed essential. Writers were asked to restrict, themselves to a list of 1850 characters "for the time being", called Toyo-Kánji, and to use these only according to the prescribed orthographic standards; of these a reduced list of 881 essential characters was required to be taught during the six years of elementary education. (...)

Among the postwar orthographic reforms was the adoption of simplified shapes for many of the characters and elements within characters. Later, the Chinese Communists independently simplified their script [cf. DeFrancis 1967], but often with different results from the Japanese, so that the characters currently used in Pekin and Tokyo have less in common today than they had at an earlier period.

How Alphabets Might Reflect Language

2010-02-02T05:30:00.000-08:00

(Edward S. Klima)

Of course, all writing systems are arbitrary to some extent, but the arbitrariness is minimal if orthographic units are correlated with the elements of the linguistic structure -- linguistic elements that the language user already knows just by the fact that he knows the language. (...) Having learned the sound forms of the words of the language -- which are themselves in essence arbitrary with respect to their meaning -- he would then have to associate with these as many arbitrary sequences of letters. Such system of writing does not capitalize on a structural characteristic of every human language, namely, that the sound forms of the thousands of words in each are all made up of a limited number of distinctive sounds. Thus there is a reduction of arbitrariness when letters, while themselves arbitrary, have a fixed relationship to the form of the word. (...)

If orthography provides a visual representation of an utterance (words, expressions, phrases), then an orthography is not sufficiently expressive if two words, which have different inherent sound forms, do not differ in their orthographic representation. We shall assume that an optimal orthography would not tolerate a situation like that in English where read (present tense) and read (past tense) or lead (meaning 'to conduct') and lead (the metal) have a different sound-form but have the same orthographic representation. Our consideration of the expressive power of a writing system need not necessarily be restricted to the sound form of the word, to the exclusion of other levels of language. Consider, for example, the general phenomena of homophony -- when two different lexical items have the same sound forms: bill 'a demand for a payment' versus bill 'the break of a bird'; meat 'flesh of an animal' versus meet 'a gathering'. The orthography is more expressive of the language if it distinguishes these forms, while identical at the sound-level, are distinct at the lexical-level. With such a distinction the orthography would be more expressive than the language is at the sound level.

Segments, Features, and Analysis by Synthesis

2010-02-02T05:29:00.000-08:00

(Kenneth N. Stevens)

In the case of some classes of segments, the characteristics of the radiated sound when the articulators are in their target positions are not sufficient to identify all the features of the segment. For example, when there is a complete closure at some point along the length of the vocal tract, as in the case of a stop consonant, no sound emerges from the mouth. Acoustic attributes that identify place of articulation for a stop consonant reside in the time interval when the articulators are moving either toward or away from the constricted configuration. Thus the stop consonant like [d] occurring before a vowel, information about the place of articulation is carried in the 40-odd msec following release of the closure, that is, after the tongue blade has shifted away from the target position for [d] and is moving toward a position appropriate for the following vowel. These acoustic cues reside in a relatively brief, transient time interval. All phonetic segments with the feature [+consonantal] [Chomsky and Halle 1968] seem to have this attribute that at least some acoustic cues for the segment reside in the rapid transition toward or away from the articulatory target and are, in a sense, nonsynchronous with the target.

Reading for Some Purpose

2010-02-02T05:28:00.000-08:00

(Eleanor J. Gibson)

I admit to one more basic assumption, about language in general this time. The psychological basis of language is abstraction. It is not just emiting sounds in a given sequence, but a conceptual ability of one or many kinds that makes language possible.

(...)

I think abstraction is accompanied in perceptual learning by a kind of opposite process of filtering. The invariant relations are abstracted and enhanced, as neurologists have suggested for auditory acuity [von Békésy 1967] by an inhibitionay process of filtering out what is irrelevant. William James spoke of abstraction as a process of "dissociation by varying concomitants."

(...)

One thing should be said about the phonological system, even in this abbreviated account. What the child learns is not a mere collection of items of information about sounds at different levels of structure. He learns, rather, a phonological rule system. In his native language, whatever it is, certain sounds may precede and follow others. Some combinations, on the other hand, are not permissible. There are typical consonant clusters that can begin a word, and typical ones that can end it.

Although acquisition of the phonological system is as yet poorly understood, acquisition of the semantic properties of language is a still greater mystery and a matter for psychological dispute. (...) Psychologists have made little progress in the study of meaning. But we know that every child acquires a lexicon - a vocabulary of the words of his language.

The third aspect of speech is its syntactic rule structure. Words from the lexicon are not strung together hit or miss, but in an agreed-upon order. The function of syntactic rules is often said to be that of linking conceptual structures with the phonological representation of them [Langacker 1967]. In any case, the syntactical rule structure is a complex one, specifying the sequential arrangement of units, types of units, and their hierarchical arrangement.

Arbitrary Nature of Language

2010-02-02T05:26:00.000-08:00

Monolingual persons take language so much for granted that they often forget its arbitrary nature and cannot distinguish words from things. Thus, primitive people often believed that putting a curse on somebody's name could actually harm his person.

(Languages and Symbolic Systems. Yuen Ren Chao; Cambridge University Press 1968).

History of linguistics

2010-02-02T05:24:00.000-08:00

The study of language used to be part of the humanities, especially when it was concerned with historical texts, the study of which constituted philology. People talked about philology long before there was such a subject as linguistics. When at the turn of the century the physiology of speech was studied instrumentaly in the laboratory, there began the study of experimental phonetics, thus bringing the study of language closer to the natural sciences. The progress of experimental phonetics, for reasons we shall see below, became stagnant for a time, or at least was not fast enough to catch up with the rapid advanves made in the survay and analysis of languages by the methods of direct study of the speakers' utterances, especially of unwritten languages, such as those of the underprivileged cultures; thus linguistics, which became a recognized branch of study in the middle of the first half of this century, has been more or less associated with the social sciences, especially with antropology. Experimental phonetics however never became obsolete and, with the advent of the electro-acoustic technique of handling speech sounds, it has acquired a vigorous new lease of life.

(Language Technology, Language and Symbolic Systems, Yuen Ren Chao)

O Advento da Fonética

2010-02-02T05:18:00.002-08:00

(Histótia da Lingüística, Joaquim Mattoso Camara Jr.)

Em meados do século XIX, a lingüística deu um grande passo com o advento de um estudo completo de fonética. Vimos que o conhecimento fonético não fora forte entre os gregos. Sua conexão com a linguagem tinha sido através da escrita. Vimos também que é a escrita que chama a atenção aos homens para a linguagem e os faz para para observá-la, uma vez que a atividade da fala é espontânea, o que a faz parecer tão natural como as outras atividades corporais, tais como o andar.

Os maravilhosos resultados obtidos pela lingüística comparativa indo-européia estimulou a investigação lingüística nos mesmos moldes no âmbito das línguas não indo-européias. Havia uma profunda impressão de que todas as línguas do mundo poderiam estar sujeitas a um estudo histórico comparativo semelhante.

(...)

O método histórico-comparativo não foi, em muitos casos, claro e firme. Em outros casos, na realidade, foi abandonado por um modo de pesquisa empírico. Os lingüistas treinados na técnica do indo-europeu não se sentiam atraídos a estudar línguas e grupos de línguas nos quais eram incapazes de aplicar seus métodos rigorosos de investigação, de modo que este estudo foi deixado a missionários e antropólogos.
Alguns deles, é verdade, tinham um certo preparo lingüístico; mas tendiam a encarar a linguagem como qualquer outro fato antropológico. Dessa forma a lingüística dissolveu-se na antropologia e perdeu o aspecto característico de uma ciência particular; que o estudo histórico da linguagem lhe houvera atribuído desde o começo do século XIX.

Com o passar do tempo, isto provou ser benéfico, porque desenvolveu uma nova abordagem liberta da visão demasiadamente estreita que considerava a lingüística como uma disciplina meramente histórico-comparativa.

(Capítulo XV - Estudos não-indo-europeus. Gramáticas comparativas. A síntese da Comparação Lingüística, Histótia da Lingüística, Joaquim Mattoso Camara Jr.)

Foi, entretanto, Forchhammer; entre os foneticistas do começo do século XX, que deu um passo à frente, de maneira mais decidida nesse campo ainda novo. Faz diferença, na realidade física e fisiológica, entre "sons vocais" (Sprachlaute) e sons da transição ou glides (Übergangslaute); acentua também que, para qualquer som vocal, há diferentes colorações sonoras (Lautschattierung). O própósito lingüístico, de acordo com ele, superpõe-se a todas estas diferenças físicas e, na língua, encontramos um grupo de sons representados por uma letra (Buchstabenlautgruppen). Em outras palavras, Forchhammer acha que as letras na linguagem escrita representam a síntese de vários sons, que são distinguidos uns dos outros pelos falantes e ouvintes comuns. Veremos agora como este conceito se encontra na base de uma nova abordagem aos sons vocais.

O primeiro estudioso a dar um passo firme nesta nova abordagem foi o polonês Jan Baudouin de Cartenay, cujo principal trabalho no fim do século XIX se intitulava 'Investigação para uma Teoria das Alternâncias Fonéticas'. Courtenay, que era professor de Lingüística na Universidade de São Petersburgo, admitia a distinção entre os sons que eram realmente emitidos e os que falantes acreditam fazê-los e os ouvintes julgam ouvir. Os primeiros se constituem no objeto da investigação da fonética; mas os últimos tinham um conteúdo lingüístico, umas vez que é por meio destes sons "pretensamente" emitidos que a comunicação se realiza. Chamou a esses sons
"intencionais" de "fonemas" em oposição àqueles emitidos realmente, que são os sons vocais da fonética. O termo "fonema" já existia na Grécia Antiga, com o significado de "enunciação" ou "voz". Este termo foi sugerido a Baudouin por seu notável discípulo
e colaborador Kruszewski, que é o autor de um tratado sobre 'Mudanças Fonéticas'.

Baudouin definia o fonema como "a idéia de um som vocal" e advogava uma análise psicológica a fim de se chegar a ele, partindo do nível da fonética e seus sons vocais. Isto quer dizer que atribuía o som vocal à física e o fonema à psicologia.

(...)

Trubetzkoy, que tinha sido aluno de Baudouin em São Petersburgo, era tanto um europeísta quanto um especialista em línguas caucasianas. Relacionado com ambos os estudos lingüísticos fez ele trabalho de valor inestimável. Já vimos como contribuiu para as novas abordagens à Lingüística Comparativa. Mas sua contribuição mais importante e notável foi sua investigação teórica a respeito do novo conceito de
"fonema", no qual, juntamente com seus companheiros, chegou a uma abordagem completamente diferente da que Baudouin apresentara.

Forçado a deixar a Rússia depois da Revolução Comunista de outubro de 1917 (era um príncipe, primo do czar), estabeleceu-se na Áustria e lecionou em Viena e Praga. Nesta última cidade foi criado um Círculo Lingüístico (1926) com ele e seus amigos e colegas russos acima mencionados e estudiosos tchecos tais como Vilhelm Mathesius, que lançou a idéia de círculo lingüístico, B. Trnka e J. Vachek; o Círculo publicou anualmente um volume das atas sob o título francês de 'Travaux du Cercle Linguistique de Prague', com artigos escritos em francês, inglês e alemão. O primeiro volume destes 'Travaux' data de 1929 e a série foi interrompida logo depois de 1941 no seu décimo volume. O nono e décimo (1939) volumes foram publicados após a morte de Trubetzkoy, em princípios de 1939, quando Viena foi invadida pelos nazistas. O novo volume é a obra póstuma de Trubetzkoy, escrita em alemão, e intitulada 'Princípios de Fonologia'.

O Círculo fez sua primeira aparição na Europa em 1928, durante o Primeiro Congresso Internacional de Lingüistas realizado em Haia. Neste congresso Jakobson, Karcevski e Trubetzkoy apresentaram uma comunicação na qual discutiam os métodos mais convenientes para uma descrição prática e completa da gramática de uma língua como ofereciam também, nessa oportunidade, o conceito de entidades lingüísticas ao qual haviam chegado em face da realidade física do discurso. Propuseram a criação de dois estudos distintos: a fonética, uma ciência natural, e a fonologia, uma parte lingüística que trata da significação dos traços fonéticos em uma língua. Essa comunicação, que é muito breve e que foi escrita por Jakobson em francês, enfatizava: 1) a necessidade de
"correlações fonológicas", isto é, do estabelecimento de um sistema de oposições de sons lingüisticamente significativos, de acordo com o princípio das oposições lingüísticas de Saussure; 2) a importância de partir dessas correlações para explicar a mudança fonética (uma idéia inteiramente nova que hoje tem sido intensamente desenvolvida, como veremos). Este segundo aspecto daquela comunicação visava a fonologia diacrônica; não foi levado em consideração, entretanto, no trabalho de Trubetzkoy, que permaneceu no nível sincrônico.

(Capítulo XXVII - O Conceito de Fonema. Baudouin de Courtenay, Saussure e o Círculo de Praga, Daniel Jones, Joaquim Mattoso Camara Jr.

The Sound Shape of Language

2010-02-02T05:18:00.001-08:00

(Roman Jakobson and Linda R. Waugh)

In the Greek philosophical literature indivisible sound units capable of forming meaningful strings were termed STOICHEIA, 'the underlying primes of sounds and letters'. The sound shape of language and correspondingly its alphabet were viewed as a joint coherent system with a limited number of discrete and interconnected formal units. This concept proved to be so persuasive that Democritus (fragment A6; cf. Diels and Wilpert) and his adherent Lucretius, in searching for an analogy which might confirm their theory of the atomic structure of the physical universe, cited STOICHEIA as the minimal components of speech.

Just as in Antiquity the term STOICHEION, used originally for linguistic elemental units, was extended to the physical world, in a similar way, but reversely, linguistic theory of the past hundred years in its quest for the ultimate constituents has appealed in turn to the model of atomic physics.

New horizons in the study of language

2010-02-02T05:16:00.000-08:00

(Noam Chomsky)

This property intrigued Galileo, who regarded the discovery of a means to communicate our "most secret thoughts to any other person with 24 little characters" as the greatest of all human inventions. The invention succeeds because it reflects the discrete infinity of the language that these characters are used to represent. Shortly
after, the authors of the Port Royal Grammar were struck by the "marvellous invention" of a means to construct from a few dozen sounds an infinity of expressions that enable us to reveal to others what we think and imagine and feel - from a contemporary standpoint, not an "invention" but no less "marvellous" as a product of biological evolution, about which virtually nothing is known, in this case.

Theories of Forgetting

2010-02-02T05:15:00.000-08:00

Forgetting was first studied in detail by Hermann Ebbinghaus (1885/1913). He carried out numerous studies with himself as the only participant. Ebbinghaus initially learned a list of nonsense syllables having little or no meaning. At various intervals thereafter, he recalled as many of the nonsense syllables as possible. He then re-learned the list. His basic measure of forgetting was the savings method, which involved seeing the reduction or saving in the number of trials during re-learning compared to original learning. Forgetting was very rapid over the first hour or so after learning, with the rate of forgetting slowing considerably thereafter (see Figure 6.12). These findings suggest that the forgetting function is approximately logarithmic.

(...)

It is often believed that forgetting is a bad thing, and that we should do everything we can to reduce forgetting. This belief is not altogether correct. We often need to update our knowledge, and it is actually helpful to forget the previous state of affairs. For example, when driving you might find it hard to remember the speed limit applying to the area through which you are driving if you had a clear recollection of different speed limits during earlier parts of the drive.

As variantes e suas conceituações

2010-02-02T05:13:00.000-08:00

(Fonética e Fonêmica, Mattoso Câmara)

O fonema, entendido como um feixe de traços distintivos, individualiza-se e ganha realidade gramatical pelo seu contraste com outros feixes em idênticos ambientes fonéticos. Não é, pois, a diferença articulatória e acústica que distingue primariamente dois fonemas, senão a possibilidade de determinarem significações distintas numa mesma situação fonética.

Compreende-se assim que um mesmo fonema possa variar amplamente, na sua realização, conforme o ambiente fonético ou as peculiaridades do sujeito falante.

(...)

pode dar-se a circunstância de que, em certas condições de ambiente fonético, uma oposição de fonemas, em geral depreensível na língua, se anula e a realização física passa a ser uma só, quer sob o aspecto de um dos fonemas, quer sob um terceiro aspecto em que só se mantém os traços comuns a ambos. É o chamado fenômeno da neutralização (al. Aufhebung; ing. neutralisation) com a presença do "arquifonema" em que se anula a oposição. Um bom exemplo em português é o contraste entre sibilantes e chiantes (/s/ e /x/, /z/ e /j/) em posição pós-vocálica (realizando-se sempre a chiante na maioria do território de língua portuguesa, ou parcialmente sempre a sibilante).

Data-Processing Inequality

2010-02-02T05:09:00.000-08:00

O teorema de processamento de dados mostra que não existe processamento algum que possa ser feito sobre algum dado de forma a melhorar as inferências que possam ser feitas sobre o dado.

teorema: (desigualdade no processamento de dados)
Se X -> Y -> Z, então I(X;Y) >= I(X;Z).

As variáveis aleatórias X, Y e Z formam uma cadeia de Markov nesta ordem (X -> Y -> Z) se a distribuição condicional de Z depende apenas de Y e é condicionalmente independente de X, ou seja, a função de densidade de probabilidade conjunta poderá ser escrita da seguinte forma:

p(x,y,z) = p(x)p(y|x)p(z|y).

Temos uma cadeia de Markov se e somente se X e Z são independentes dado Y. O fato de ser uma cadeia de Markov implica em independencia condicional pois:
p(x,z|y) = p(x,y,z) / p(y)
= p(x,y)p(z|y) / p(y)
= p(x|y)p(z|y).

Para demostrar o teorema de processamento de dados, basta expandir a informação mútua das seguintes formas:
I(X;Y,Z) = I(X;Z) + I(X;Y|Z)
= I(X;Y) + I(X;Z|Y)
Por ser uma cadeia de Markov, X e Z são condicionalmente independentes, dado Y, e assim I(X;Z|Y) = 0. Como I(X;Y|Z) >= 0, temos:
I(X;Y) >= I(X;Z).

A igualdade será apenas quando I(X;Y|Z) = 0 (i.e., quando também for uma cadeia de Markov X -> Z -> Y). De forma similar mostra-se que I(Y;Z) >= I(X;Z).

Se Z = g(Y), X -> Y -> g(Y) forma uma cadeia de Markov, então I(X;Y) >= I(X;g(Y)). Uma função do dado Y não consegue melhorar a informação a cerca de X.

Sendo X o sinal transmitido, como Y o sinal recebido. Não existe processamento que possa ser feito sobre Y de forma a obter mais informação a cerca de X. Qualquer processamento que seja feito sobre Y resultará em um Z cuja a informação mútua com X será igual ou menor à informação mútua de Y com X.

Contemporary Views

2010-02-02T05:03:00.000-08:00

(Writing Systems, An introduction to their linguistic analysis, Florian Coulmas)

Although there is plenty of evidence that, literate cultures, writing intrudes into linguistic behaviour of people and that without writing many languages would not be what they are, the notion that writing is an active agent of language is unpalatable to many linguists for a number of reasons. One is that in modern linguistic languages are stripped of their historical dimension. Although the obvious fact that language change in the course of time is acknowledged, the possibility that their nature may be affected by external factors such as writing is strictly denied, allegedly on the grounds that writing could not possibly have exercised any influence on the faculty of language because it is too recent an invention. The oldest records reach back a bit more than ten thousand years at best, while language must have evolved hundreds of thousands of years ago. Diachronic linguistics is essentially unhistorical, because, as a defining capacity of the human race, language is not supposed to change by virtue of humanly contrived technology. There are no highly or less highly developed languages. This is a primitive of linguistics. Artifacts and technologies, such as writing, for example, are granted the potential to change the environment, but not humanity itself. Since language is conceived as an essential part of human nature, while writing is a mere technology, the effects of writing on language and by implication the complexities of their interrelationship remain largely unexplored.

Scholars in the language sciences who do believe that the invention or discovery of writing does make a difference, both with respect to what language is and how we think about it, are in a minority. Linguistic orthodoxy happily concurs with Ferdinand de Saussure's apodictic statement that made Aristotelian surrogationalism a cornerstone of modern linguistics:

"Language and writing are two distinct system of signs; the second exists for the sole purpose of representing the first. The linguistic object is not both the written and the spoken forms of words; the spoken forms alone constitute the object." (Saussure 1959:23)

Saussure, Ferdinand de. 1959. Course in General Linguistics. Translated by Wade Baskin. New York: The Philosophical Library.

(...)

Since the majority of languages of the world are unwritten, it is only prudent to ignore writing when studying language. However, this argument is not as sound as it seems. For, if all languages are of kind it follows that if some language are writable all languages are, and since writing is undeniably not the same as language, it is a legitimate question how the two relate to each other. Two questions linguistics should not sidestep are: 'what happens when a language is written down, (1) in terms of linguistic description, and (2) in terms of linguistic evolution?' As a matter of fact, linguistic never study any language without recording speech and writing it down. This alone is a compelling reason for studying writing instead of assuming that writing, whose essential properties are so radically different from speech, can be ignored in the research process.

(...)

Since Saussure, grammatical theory has undergone revolutionary changes, but the central concept of relating sound to meaning in a structured way has remained the same. Saussure's model of the linguistic sign still captures the main point. Sound in language has three aspects, which he distinguishes: physical (sound waves), physiological (audition and phonation) and psychological, that is, sounds as abstract units, which he calls 'sound images' (images acoustiques).

"The linguistic sign unites a concept and a sound image. The latter is not the material sound, a purely physical thing, but the psychological imprint of the sound, the impression that it makes on our senses." (Saussure 1956:66)

(...)

"Our conception of language is deeply influenced by a long tradition of analyzing only written language, and that modern linguistic theory, including psycholinguistics and sociolinguistics, approaches the structure and mechanism of spoken language with a conceptual apparatus, which - upon closer scrutiny - turns out to be more apt for written language in surprisingly many fundamental aspects" (Linell 1982: 1).

Linell, Per. 1982. The writing Language Bias in Linguistics. Linköping: Linköping University, Department of Communication Studies.

(...) Faber (1992: 110) interprets the observation that many speakers cannot divide words into phonological segments 'unless they have received explicit instruction in such segmentation comparable to that involved in teaching an alphabetic writing system' as evidence that historical segmentation ability was a consequence of alphabetic writing, not a prerequisite. Various sounds such as diphthongs and prenasalized consonants, which in alphabetic writing are represented by sequences of letters, cannot realistically be conceived as isolated steady units.

(...) David Olson (1994) stresses the point that the concept of the word as distinct unit is a by-product of literacy acquisition. Morphology, the study of words and their parts, is deeply imbued with notions of literate 'word processing', such as 'lexical entry', for example. 'Lexicon' itself is such a term. A lexicon is a list of isolated words, a kind of usage that does not occur naturally in speech. The word is an artificial entity in another sense as well. It is basically the kind of unit that is listed in a dictionary and thus not necessarily the same in all languages.

Olson, David. 1994. The World on Paper. Cambridge: Cambridge University Press.

Static entities like the stock of words, sentences and written texts are alien to the spoken language where meaning is constituted in the act of speaking, bound to situation, speaker, context, the interaction history of speaker and listener, and so on. (...) As Olson demonstrates at length, this terminology is not fortuitous but speaks of the fact that linguistics is grounded in written language. Since linguistics is concerned with 'natural language' while writing is an artifact, this is difficult to openly integrate into linguistic theory, which, as a result, is characterized by scriptism, which has been defined as "the tendency of linguists to base their analyses on writing-induced concepts such as phoneme, word, literal meaning and sentence, while at the same time subscribing to the principle of the primacy of speech for linguistic inquiry." (Coulmas 1996:455)

As Olson sees it, linguists are in this respect representative of literate society at large where writing provides the model for speech, rather than the other way around. We pronounce as we spell, we judge our utterances against the yardstick of written sentences and qualify as ellipsis, anacoluthia, reduction, false start and so on those which do not conform to these patters. The literal meaning of a sentence is basic. Other meanings are taken to be derived from it. To a scholar who, like Olson, looks at language as something to be learned, such a conception, perhaps, comes quite naturally because it is the written form of language that is made the object of instruction, memorization and testing. As an institution, the school instils into the collective mind the primacy of writing. In contrast, those who prefer to look at language as a natural capacity tend to insist on the primacy of speech. These seemingly irreconcilable positions reflect the two sides of language, the acquired and the innate. Since on human being exists as purely natural creature, both can be dissociated only in theory. This is the deeper meaning of Olson's notion that writing is a model of speech.

Consonants and vowels

2010-02-02T04:59:00.000-08:00

(Writing Systems, An introduction to their linguistic analysis; Florian Coulmas)

Take care of the sense, and the sound will take care of themselves.
Lewis Carroll

It is a recognized fact, and has been for millennia, that there are two complementary classes of speech sounds, consonants and vowels. Segmentalism, we noted in the previous chapter, is a view of language that treats both classes exactly alike, inspired to do so, perhaps, by interpreting the Graeco-Latin alphabet as an iconic map of speech sounds where letters order represents the sequence of articulated sounds. As a matter of principle, letters for vowels and consonants are assigned equal space in writing system derived from the Greek alphabet, and as a class V letters are indistinguishable in form from C letters. Indeed, the equation of both is usually quoted as the crucial accomplishment of Greek writing. Yet, there are some conspicuous differences between vowels and consonants.

Differences between consonants and vowels

(...) vowels have syllabic status, consonants usually do not. Certain consonants such as /n/, /l/ and /w/ are syllabic in some languages and, on the other hand, there are also non-syllabic vowels, but generally speaking syllabicity is more closely associated with vowels than with consonants. Vowels rather than consonants bear stress accent, pitch and tone.

Another difference between the two classes of speech sounds is seen in their relative volatility. On the stage of sound change vowels play the leading roles, while consonants are usually assigned walk-on parts. In the following examples of the Great Vowel Shift from Old English (OE) to Modern English (ME) in the fifteenth century, the vowels changed, but the consonants were preserved.

Vowels are also more prone to get changed in paradigmatic derivations such as 'foot', 'feet', 'tooth', 'teeth', 'mouse', 'mice', 'breathe', 'breath', 'sit', 'sat'. Consonants from the skeleton of these words, vowels provide the flesh. (...) To be sure, consonants too change over time, but vowels are of their nature more unstable, especially long vowels, which often end up being diphthongs, as in 'house'.

Consonants and vowels differ in production and sound quality. Vowels are pure sing-song that opera singers can use when practising their voice, a e i u ae o. Vowels are more likely than consonants to be uttered in isolation and to form words and interjections. In vowels the air stream released from the lungs is uninhibited, differences between them resulting from modifying the shape and length of the resonant cavities that the sound passes through. By contrast, consonants are hisses, hums, buzzes and puffs of air forced through the relatively constricted, or completely closed, vocal tract. They are produced by moving the tongue, the most important of the speech organs, around the mouth, obstructing the flow of air in various ways or bringing it to a stop by closing the lips. Though consonants cab be produced in isolation, ssssss, krrrr, they are typically articulated in combination with a vowel, pooh, ugh.

Yet another difference between consonants and vowels concerns their distribution and function in different languages. Most languages have a clear division of labour between vowels and consonants: syllables begin with consonants and end with vowels; clusters are rare. These are general tendencies, but there are marked differences on other levels of the language system. Some languages rely more heavily on consonants than others.