Wednesday, October 19, 2005

The Language of Causation

One of the great verbal tricksters in English is the word cause, especially when used as a noun with the definite article the, as in a sentence like (1), which is clearly false.

(1) Smoking is the cause of lung cancer.
Even occurrences of the indefinite article a when it occurs in a sentence like (2) are problematic. This sentence suggests that smoking all by itself can cause cancer which is surely false..
(2) Smoking is a cause of cancer.
The same is true of generic verb occurrences of cause, as in (3).
(3) Smoking causes lung cancer.
If you are head of a tobacco company being sued you will want people to make claims like (1),(2), and (3) since none of them stands any chance of being true. Though they are used very frequently, such claims as these are very easily falsified because every event or state of affairs will normally have multiple causes. In the case of smoking, ones genetic make up has a bearing on whether or not a person will get lung cancer. Environmental factors are likely to play a role as well.

What makes generic claims so tricky is that we readily assent to a claim like (4) even though we know that more than half of the lion population, namely the females and the cubs, don't have manes.

(4) Lions have manes.
Similarly, even though (3) makes the very strong claim that smoking alone can cause cancer, it is consistent with some people being life-long smokers and never contracting lung cancer. That is, it is consistent with (6).
(6) Smoking doesn't always cause lung cancer.
This is the beauty of generic claims -- they make very strong (because highly general) claims but they aren't falsified by counterexamples. They live a good life, as the lives of propositions go.

I ran across a number of claims containing "not the cause of" such as

(7) HIV Is Not the Cause of AIDS
(8) Cosmic Rays Are Not the Cause of Climate Change, Scientists Say
(9) Income inequality is not the cause of this nation's social problems.
(10) Gun ownership is not the cause of America's high murder rate.
The last two come from the same web site and are quite comical. Only an idiot would claim that income inequality is THE cause of the nation's social problems or even worse that gun ownership is THE cause of America's high murder rate. Obviously, for a murder to occur, the gun owner will have to load the gun, take the safety off, point it accurately at the intended victim, and pull the trigger and do all of this with the intent to kill (at least for first degree murder) so we know that gun ownership simpliciter does not cause murder. Moreover, there are causes of murder not involving guns such as strangling someone with ones hands or using a garrote or knifing someone to death (cf the O. J. Simpson murder trial) or running them over with a car or whacking them with a baseball bat and etc.

The eighth claim is more interesting. The claim at issue is this one

In July 2003, astrophysicist Nir Shaviv and geologist Jan Veizer wrote in GSA Today that they had established a correlation between cosmic rays and temperature evolution over hundreds of millions of years. They also claimed that current global warming is not primarily caused by human emissions of carbon dioxide. Their findings have been widely reported in international news media.
Notice first that (8) is actually the headline for a news release and was almost surely composed by a PR guy/gal, not an actual scientist. Moreover, the passage quoted does not provide support for the idea that cosmic rays are THE cause of climate change. All the scientists seem to have established, if they established anything, was "a correlation between cosmic rays and temperature evolution over hundreds of millions of years. [emphasis added]" I don't know whether these scientists said anything like "current global warming is not primarily caused by human emissions of carbon dioxide [emphasis added]" but this kind of claim is of interest as well. Scientists who ought to know better say things like "the primary reason for X is Y" or "the real reason for X is Y." I suspect I have used both in the past. These are naughty claims. When you have a set of causes of some event Y, how do you establish one of them as the primary one? [At this point I might have gone into Aristotle's four causes but I have been warned against having overlong blogs.]

As sentient beings who are proud of our mental possessions, we tend to attach great importance to our decisions and actions as event movers and shakers. But there are always other factors. When, as in the case of cosmic rays, there are no human sentient beings involved religious people will sometimes evoke God's choosing to subject us to his/her wrath by, say, raining cosmic rays onto us. Jerry Falwell once said, for instance, "AIDS is not just God's punishment for homosexuals; it is God's punishment for the society that tolerates homosexuals." Who knew? I thought it was a virus. We have to expect that from God, I suppose, for as Archibald McLeish wrote in his play JB "If God is God, God is not good. If God is good, God is not God."

Right now, plaintiff's lawyers are going after Merck and its product, Vioxx. In this and in all other liability cases, I suspect, the issue of who caused what and the relative importance of different causes lie at the heart of the suit. I picked up the following collection of sentences from the article linked to the title of this blog. There is not just the headline (11), but also (12)-(17), for you to entertain yourself with.

(11) Cardiologist: Vioxx did not cause postal worker's heart attack.
(12) A postal worker who suffered a heart attack had a buildup of plaque in his arteries that was not caused by the since-withdrawn painkiller Vioxx,...
(13) ...only minimal plaque buildup is needed to cause the "small, modest" heart attack...
(14) Humeston, 60, is suing Vioxx maker Merck & Co., blaming his heart attack on intermittent use of the drug over two months.
(15) The plaque broke off, causing Humeston's heart attack, he said.
(16) ... Tyberg testified that Vioxx does not cause plaque buildup.
(17) "Did Mr. Humeston have sufficient plaque in his arteries to cause a heart attack?" Sullivan asked.
This gives you a taste of how loosely "cause" and what we might call "causal words" such as "blame," as in "X blames Y on Z," were used in the article and in court. Not a single one of these uses is intellectually respectable. Even (15) doesn't attribute the attack to what Aristotle might have called "the final cause" -- that would have been, the obstruction of blood flow. It could be that narrowed arteries at the point the clot lodged itself played a role as well. Perhaps the best guidance for us is to drop the word "cause" from our vocabularies. The chances are that nothing good will come from their use.

Tweet This!

Monday, October 17, 2005

The Origin of Language

Two of the great mysteries of human development are

(I) How did spoken language originate?
(II) How do children learn language so quickly?
The first of these two questions is rarely discussed by contemporary linguists for the simple reason that we don't have a clue how language came into being. We know from an evolutionary standpoint that the difference between primate and human vocal apparati was critical to language development but as was noted by Carl Zimmer, who is quoted at a religious "Origin of Language" site as saying in his book Evolution:
No one knows the exact chronology of this evolution, because language leaves precious few traces on the human skeleton. The voice box is a flimsy piece of cartilage that rots away. It is suspended from a slender C-shaped bone called a hyoid, but the ravages of time usually destroy the hyoid too.
Typically, when scientists don't have an answer to a question they keep their yaps shut. And we do that with one exception and that consists of comparing such things as the human "voice box" Zimmer refers to and the primate vocal apparati and comparing parts of the human brain known to be involved in language processing with corresponding parts of primate brains. That is, we can hope to say what it was that evolved that allowed languages to come into being, but we cannot say how and when it did so.

So, what linguists do is focus on the second question. It is a great mystery how it is that children learn language in a relatively brief period -- from something like the age of 2 to the age of 12 or so -- and they do this without being taught. Parents and others provide models and parents and others also provide corrections though frequently the latter are totally lost on young children for they are seen not as corrections of linguistic form, a notion that is quite abstract, but as a denial of the truth of what they say. Some very good advice is given at a Kid Source site where it is said:

How can I help a child pronounce words correctly?

* By setting a good example. Don't interrupt or constantly correct the child. Don't let anyone tease or mock (including friends or relatives). Instead, present a good model. Use the misarticulated word correctly with emphasis. If the child says, "That's a big wabbit," you say "Yes, that is a big rabbit. A big white rabbit. Would you like to have a rabbit?"
The keys are confirming the truth of what they child has said while providing a correct model. And what is said about pronunciation holds for corrections of grammar. We had a picture of a pig in our linguistic offices for years which said, "Teaching a pig to speak annoys the pig and wastes your time." The same is sometimes true of correcting a young child's speech.

One of the truly unfortunately developments in linguistics in my opinion has been Chomsky's focus on what he has sometimes called "the language organ." In an Q and A with a BBC interviewer in 1996, Chomsky says, speaking of the problem linguists face:

So, the main goal was: find the actual rules of language. Then the next goal would be: explain how they got there. Well, to explain how they got there you have to go back and ask: what's the initial state of the language faculty? What's its initial design, presumably common to the species, because we're not adapted to learn one language or another? So, what is the initial design of the common language faculty that enables it to take these highly intricate, closely articulated, delicately structured forms very rapidly on the basis of minimal interaction with the environment? It's a typical problem of growth -- you know, of growth of organs -- in this case the growth of the language organ.
I shudder every time I think of Chomsky's "language organ."

As I said in my last blog, when I showed up at M.I.T. I was disposed not to take very seriously Chomsky's talk about our innate language faculty. Why did/do I feel this way?

In my opinion, the approach of Chomsky to the second question asked at the beginning of this blog about how children learn language is little different from the sort of answer that Christian and Jewish religious fundamentalists give to the first question I asked above. At the religious Origin of Language site I cited earlier one reads:

When God created the first human beings—Adam and Eve—He created them in His own image (Genesis 1:26-27). This likeness unquestionably included the ability to engage in intelligible speech via human language. In fact, God spoke to them from the very beginning of their existence as humans (Genesis 1:28-30). Hence, they possessed the ability to understand verbal communication—and to speak themselves!

God gave very specific instructions to the man before the woman was even created (Genesis 2:15-17). Adam gave names to the animals before the creation of Eve (Genesis 2:19-20). Since both the man and the woman were created on the sixth day, the creation of the man preceded the creation of the woman by only hours. So, Adam had the ability to speak on the very day that he was brought into existence!

Notice how easy it is for the people responsible for the Origin of Language site to answer this amazingly difficult question. All they have to do is quote the Bible.

If science were this easy, anyone who can read could do it. It is hard to figure out what the human vocal and auditory apparati consist of and how they might have evolved and how the brain evolved to provide the capacity for humans to learn the languages we speak today, as well as learn how to use them. When God created Adam he is said to have been given the capacity to speak instantly. That is, he was not only provided with what Chomsky calls "the language organ," he was given a specific, full blown language, with its sounds, its morphemes, its words, and the grammatical rules that allowed Adam to speak to Eve (also outfitted with this stuff) the moment she was created. This is a point of view that is staggeringly simplistic.

I feel the same way about Chomsky's answer to the second question with which we began. How do we learn languages? Easy, Chomsky says. We take our "minimal environmental" data and let our language organ go to work on it. Of course, linguists who try to specify the features of this language organ or, what is the same thing, who try to specify the features of universal grammar face a very difficult task. That takes very hard work. But at a fundamental level, Chomsky's answer to the question of how children learn language is as simplistic as is the Bible quoter's answer to the question of where human language came from.

Notice Well: I am not making fun of religion. There are numerous religious points of view that do not involve providing simplistic answers to difficult questions. I learned years ago that I am not an atheist per se. I was raised as a Southern Baptist and it is those teachings I reject. That makes me a fundamentalist atheist I suppose. As for other religious points of view and there are very many including religious views of very smart, sophisticated people I have nothing to say. They don't interest me. As I suggested, when a scientist knows he can't answer a question, his best option is to stay silent.

Tweet This!

Thursday, October 13, 2005

Language Universals and Language Learning

This is a long post but it is leading up to something pretty wild and wooly.

I made reference tomy Rice professor, Trenton Wann, in my blog "What Constitutes a Human?" One of the things he taught me while I was at Rice (in class and in a lot of private chats thanks to his willingness to suffer my questions and arguments) was a deep suspicion of nativistic theories of human abilities. Before going to M.I.T., I knew that I would be running into the center of nativism in linguistics, for Noam Chomsky was arguing that we humans have an innate knowledge of language that takes the form of universal grammar. Naturally, I would be suspicious of that claim and I did not make myself popular when I expressed my doubts. One student called me "Stupid." That stung but I knew that she was being stunningly naive about what must be proved to prove that a human ability is innate. I didn't say so because I had not yet been fully admitted into the graduate program and she was a star student.

Chomsky did not hold that we speakers of English are born knowing English. That would be a very silly claim to make and Chomsky is not a silly man. It is evident to anyone who looks that the language anyone learns is a function of what language is spoken at home and, when it is different from the "official" language of the society, the language that is taught at school. Thus in the Arab world, everyone learns a regional variety of Arabic at home, and at school kids learn the standard language, which is more or less the language of the Koran. (I am not sure about the actual relationship between the standard language and the language of the Koran.)

What Chomsky meant by our having an innate knowledge of language is that we have an innate universal grammar. Against such a notion, it has long been recognized that languages can be very different in their "surface" structure, that is can be different as to such things as the specific sounds employed, how these sounds are organized into word bits called "morphemes", how these morphemes are combined to form words, and how words are combined to form phrases and sentences. There are languages like Vietnamese that exhibit very little morphology (noncomplex words not containing morphemes marking tense (walk vs. walked, number (boy vs. boys), gender (prince vs princess), case (she vs. her), etc. Such languages tend to exhibit strict word order so that case relations (subject vs indirect object vs direct object, etc.) are indicated clearly (Bill loves Sue doesn't mean Sue loves Bill). Other things marked by morphemes are provided by context in such languages. Though I have used English for all of my examples, English also exhibits very little morphology. Latin, as those who were, like me, forcibly introduced to it will know, exhibits a good deal of morphology and therefore can allow greater freedom in word and phrase order (such differences often reflecting other aspects of interpretation besides literal or conventional meaning).

There are polysynthetic languages like Inuktitut (an Eskimo language) in which whole sentences can be expressed in a single word thanks to the fact that many morphemes can attach to the word root that would in Vietnamese be individual words. This creates some serious problems for the notion of a universal grammar.

Most languages seem to distinguish the subject from predicate (verbs, both main and helping verbs, and direct and indirect objects and adverbials (prepositional phrases and one word adverbs such as to the store and there.) However, there is trouble in paradise. There are languages that vary substantively in the order of main elements. We have the SVO languages (subject, verb, object, as in English), SOV languages (Japanese), which though different from the English pattern, maintains the subject-predictate structure, VOS (a relatively rare type, but found in Malagassy, and also a language preserving the subject predicate distinction), and VSO (a very rare type, but found in Hixkaryana, which messes with the idea that every language exhibits the subject-predicate structure). And, to screw up the works some more, there are ergative languages like Basque (itself a total mystery as to its relationship to other languages) that treat the subject of an intransitive very like sleep and the object of a transitive verb like kiss the same way morphologically. The best web page I found in a quick Google search on language differences and similarities is one by Paul Hagstrom.

Despite all this variation there do seem to be some language universals which can be seen as evidence of the existence of universal grammar. As the site done by Hagstrom notes, many are implicational, stating for instance that if a language has property p, it will also have property q. So, if a language has nasal vowels (the vowel in the English word can) as part of its basic repertory of vowels (not true of English), it will also have nonnasal vowels (French is an example). But in general, to salvage the notion of universal grammar one must look to quite abstract language properties that are shared.

There is another factor to consider in evaluating whether there is a universal grammar that we are born with and this is that there seems to be a differential ability in learning languages between children (from 2 to12, roughly) and adults. Adults can clearly achieve great fluency in learning languages and some would argue that the difference has less to do with age, which equates to brain maturation and structural changes, than social, psychological, and educational factors (immersion vs learning in classes). For an overview, check out this site by
Ji-yeon Kook. I have significantly departed from my areas of expertise and won't take a position on this issue. However, it has been reliably reported that fluent nonnative speakers don't react instinctively to cursing they way native speakers do and that they are not as clear as to what is and is not a grammatical sentence in the dialect they share with native speakers.

One can also take the position that we have no language-specific innate cognitive capacity but rather we are born with a limited vocal apparatus that restricts us to a set of sounds that we can make with some being more easily made than others (stop consonants vs. others and nonnasal vowels vs. nasal vowels) and a powerful perceptual-cum-conceptual apparatus that allows us to learn that events exist that involve various sorts of relationships among them, such things as simple relations (John is tall or John kissed Mary involving an entity and a property or an actor, an action, and an acted upon) to complex causal connections (John killed the ant being cognitively parsable as John acted in such a way as to cause the ant to come to die). It could be argued that we bring that conceptual apparatus plus our developing capacity to master sound production plus our perceptual (visual and auditory) apparati to bear on language learning. (I know it is uncouth to use "plus" but I don't really care.) I would take the position that if we have these various nonlinguistic learning abilities they must themselves be innate to at least some degree for saying that we are born with a blank slate and we learn to learn before or as we learn things seems profoundly improbable.

There is still another possibility, of course, and that is that there are some language universals and there is an independent cognitive apparatus, both being innate, that we bring to bear on learning the language data we are exposed to. Nothing complex is ever just one way or the other contrary to what True Believers tend to think.

What I am going to assume here is that we do not start off with a blank slate upon which experience writes its lessons but that we have some sort of innate abilities that facilitates our learning languages. It is clear by now that though primates can be taught to communicate using symbolic language, none has ever acquired anything like a human language, namely a language that allows the embedding of one sentence inside another (I saw the boy who left) along with many other complex linguistic properties. However, the fact that they can communicate to some degree using symbolic language and can even, or so it is reported, learn some such symbols on their own does show that primates do enjoy some sort of very limited cognitive/linguistic abilities like ours. But, in contrast, human children not only learn language on their own (parents and others provide valuable input but don't teach their children their language -- they wouldn't know how) and vastly quicker than primates learn what they learn, they can learn several languages at once even when one is American Sign Language or some other sign language used by those who are hearing impaired (if that is the current PC designation.)

This is a long post, as many of mine have been, but I won't apologize for that. Let me add just one more thing. To prove that there are linguistic universals, there must be some property of one or more languages that cannot be learned from experience (from experiencing the primary language data they are exposed to and does not correspond to an independently needed cognitive ability (including thinking, memory, and all the rest). That is tough to do. But it is absolutely worth trying to do it.

Tweet This!

Monday, October 10, 2005

What Constitutes a Human?

While I was an undergraduate at Rice, sometime between 1958 and 1961, a wonderful psych professor, Trenton Wann, who influenced me more than any other single person, had us read a novel You shall Know Them by Vercors in translation from the French original. Some years later I found a copy but now it is lost. But it raises the question of what are the criteria by means of which we recognize an entity as a human, an issue that is raised by the abortion issue.

The basic plot is that evidence of a potential "missing link" is found in Africa and the Brits send an expedition consisting of a variety of different sorts of people including a theologian, I believe, and a reporter who serves as the novel's protagonist. Their mission is to find extant members of the group and study them. They are found and are observed. In the meantime the reporter discovers that a corporation, possibly French, plans to exploit these entities as slave labor so he impregnates a female. My memory is a bit hazy but I believe he brings the pregnant entity back with him and when the child is born he kills it and calls the police claiming he had committed a murder. He thereby creates the dilemma for himself that if he gets what he wants -- a determination that the child is a human and that therefore he has, in fact, committed murder -- then he will have to go to prison but the newly discovered people cannot be exploited as slave labor since they would have to be recognized as sufficiently human to make it morally repugnant. If he doesn't get what he wants, he stays out of prison but the corporation would be within its rights to exploit these nonhuman entities as it wishes (pace PETA).

The British parliament takes up the issue as to whether or not these entities are human and they are given access to all of the expedition's information. They decide that they are sufficiently human to warrant being called human (just as Neanderthals are called "human") and their reason was that in smoking their meat briefly before eating it -- too briefly either to cook it or give it a smoke flavor (I believe) -- they were exhibiting ceremonial behavior and that is a characteristic shared by all humans but no animal species. I open to you what you would look for in a newly discovered "missing link" set of beings as evidence that they are or are not human.

Notice how different this question is from the question whether the human foetus is a human. There is no question that a born child would meet our conditions for something being a human being though given how limited newborns are they wouldn't present much evidence. They don't talk, can't sit up to say nothing of being able to stand up or walk. The human foetus is clearly human, that is it has the quality of being human (adjective use). But it doesn't follow from that that the foetus is A human (noun use), a fully human entity, at least early on in the gestation period -- certainly not when it is a zygote or two-celled entity or four-celled entity, etc. You can take a look at a one month old foetus to determine for yourself whether you think it looks like a human.

Religious people, especially Christians, seem most firmly to believe that abortion is immoral and that their basis for believing his would be the Bible. I found a very silly site that gives ten Bible reasons for believing this. After reading the first two "bible reasons", which came down to the fact that the Bible uses "babe" and "infant" not "foetus" in reference to the foetus, I gave up reading this nutty site. I have better things to do. Obviously the Bible would not use "foetus" because the Bible was written in Hebrew, Aramaic, as well as in Greek. And this dude was reading an English language bible.

In any event, this blog site will not accept any religious text, including the Hebrew or Christian Bibles, the Koran, or whatever as an authority on any issue. The problem with arguments that appeal to such texts is that we don't all believe in the same texts so there will be no common ground upon which all of our arguments can stand. In any event, any moral principle that cannot be defended on nonreligious grounds isn't worth the pixels it takes to print it on a computer screen.

Tweet This!

Saturday, October 08, 2005

On Reasoning about Abortion

Full Metal Attorney suggested in a comment that I visit his site for a dissertation on the various arguments given for and against the pro-life and pro-choice positions. You may not get through it all but if you go deeply enough into his discussion, you will, I think, see why various arguments for and against the pro-life and pro-choice sides will fail to be persuasive to the other side. However, that does not make the debate pointless. There are always young minds that are coming to the issue without having already prejudged the issue and they may benefit from such discussions.

The fact is that this issue is unresolvable. Our best hope, expressed in several posts is that as people receive more information about birth control and the need for protection against sexually transmitted diseases, the arguments may be mooted by the need for abortion waning over time. The terrible irony is that the nitwit Right Wing Christians who most oppose abortions also frequently oppose sex education, the free availability of condoms, etc. "Stupid is and stupid does," as Forest's mom said.

In reading Full Metal's post, it struck me that there is an ambiguity in the word "human" that has a significant bearing on arguments for and against the pro-life and pro-choice positions and this is that the claim that the foetus is human does not entail that the foetus is a human. NB: At this point I make a blunder saying "Linguistically, this is the distinction between the mass noun use of "human" and the count noun use" when I should have said, "Linguistically, this is the distinction between an adjectival use, treating human as a quality and a count noun use." No one disputes that the foetus is human in the sense that it consists of human DNA as opposed to chimpanzee or any other specie's DNA. However, the human zygote which is the single cell entity that results from fertilization of a woman's egg may have human DNA but is arguably still not a human.

The passage in Full Metal Attorney (where "P" stands for a premise) that brought this issue to mind is:

P1: Murder (the unjustifiable destruction of a human) is wrong.
P2: A woman’s body is her own business and no one else’s.
P3: Abortion destroys a fetus/embryo/zygote.
Sub-conclusion 1: If a fetus/embryo/zygote is human, then abortion is murder and is wrong.
Sub-conclusion 2: If a fetus/embryo/zygote is not human, then abortion is acting on a woman’s body and the state does not have a right to prohibit it.
Those who are adamantly opposed to the killing of innocent humans, will not automatically be opposed to abortions since in the phrase "the killing of innocent humans," "humans" is being used as a count noun. The debate about when, during the process of gestation, abortions will and will not be permitted is about when we can sensibly speak of the foetus as being a human. My linguistic sensibilities are offended by the notion that anything less than the product of birth would give us a human though clearly the foetus increasingly looks like a human as it develops. But, saying that something looks like a human doesn't mean that it is a human. Years ago, a friend who had hunted bears earlier on in his life said that a skinned bear "looks like a human," but I did not instantly come to the conclusion that skinned bears are human. In short, the fact that the foetus looks like a human does not entail that it is. The decision to say that the foetus at time t becomes a human will have to be based on a better argument.

[NB: I have elided a paragraph because revising it in the light of the earlier mentioned blunder would be too much trouble. As the saying goes, "to err is human," and so I must be human.]

Several commenters stated baldly that abortion is murder. This is a really quite childish thing to say. The word "murder" is a technical term in the law and any ordinary language use of it, as in a statement like "abortion is murder," has legal implications. It would not be childish to say that abortion is the murder a human and as a result, anyone who performs one and who allows one to be performed on her body should be prosecuted for murder and for conspiracy to commit murder and since conspiracy to commit murder is a felony, both parties should in fact be charged with committing felony murder and should thus be death eligible. If you don't have the balls to say explicit things like I have just said, then please don't go around saying "abortion is murder." As I said, it is childish.

Tweet This!