Sportive Thoughts

I've previously written on my suspicion that linguistic factors influence test-scores and that genes have a direct effect on linguistic ability. Now I want to take these musings a step further, and ask whether an open class of genes-- call them i-genes --are responsible for achievement in the classroom.

Before I start, I should clarify that I do not see a causal connection between i-genes and a child's first language. To argue that there is such a connection is ludicrous; very obviously, children are not programmed by genetics to speak a given language. We know this based on what we already know about language acquisition: it occurs uniformly in a stage spanning from 12months through 4 years. All humans learn their first language during this period from their home environment. However they do it, they do it at the same time and nearly the same pace without regards as to which language they learn. Local, regional, and global differences in language use only become an issue as the child grows up and demonstrate proficiency to adults who may use a slightly different language.

With that said, while there is no way i-genes can select which language a child speaks, it is plausible if not discoverable that i-genes can show a child's ability to speak any language to be greatly improved. Thus the central claim here is i-genes are directly related to basic cognitive ability.

That much is undoubtedly true. We see children inherit very many traits from their parents, and if we do not question whether height or eye-color is a direct expression of inherited genetic material, then we must accept that some amount of cognitive ability is also inherited. The major questions are how much? and does inheritance explain the existence of racially based student achievement gaps?

How Much Intelligence is Inherited?

The science of genetic inheritance is very old, and we need not learn much of it to understand how parents' intelligence can be passed on to sons and daughters. In the simplest terms, we can imagine four different i-genes: a dominant ("S") and recessive ("s") smart i-gene, and a dominant ("D") and recessive ("d") dumb i-gene.

Let's say Adam is quite average, with the following i-genetic makeup:

sd

Nothing spectacular about Adam; he's got a recessive version of both genes. Now let's consider his partner Eve, who has a different i-gene expression:

SD

Eve is an exact average with a dominant smart gene and a dominant dumb gene. Now we can determine all the possible inheritances of Adam and Eve's children:

Ss, Sd, Ds, Dd

Here we can see, with our oversimplified version of things, that Adam and Eve will have a 50%-50% shot at intelligent children.

Because Adam and Eve's prospects are so simplified, they more or less represent what to expect from the children of very average people. Now let's consider yourself, you very smart person:

SS + SS = SS, SS, SS, SS (100% smart)

SS + SD = SS, SD, SS, SD (50% smart, 50% average)

SS + Ss = SS, Ss, SS, Ss (100% smart)

SS + Sd = SS, Sd, SS, Sd (100% smart)

SS + ss = Ss, Ss, Ss, Ss (100% smart)

SS + DD = SD, SD, SD, SD (100% average)

SS + Dd = SD, Sd, SD, Sd (50% smart, 50% average)

SS + Ds = Sd, Ss, SD, Ss (75% smart, 25% average)

SS + dd = Sd, Sd, Sd, Sd (100% smart)

So it looks like, even if your partner is extremely dumb, the expectation is that your children will be average (assuming the interaction between two different, dominant traits is the exact average). According to our simplified model, you won't have dumb children.

Now, it turns out that the results are analogus in case you are extremely dumb yourself. If your i-gene pair is "DD", then the highest intelligence your children will show is "SD", or the exact average. More than half your partners will give you children as dumb as you; only one special someone is guaranteed to give you the averge child you hope for. Suddenly things look quite a bit less optimistic.

What does all this mean?

At best it means that if i-genes are simply passed down, then a basic intelligence level will be kept in the family until it is bred out *unless* a couple starts at average to begin with (like Adam and Eve). In other words, at face value it looks like the prospects of improving children's intelligence relies on selecting for  "smart", dominant i-genes. That means we are going to have trouble justifying efforts to close any achievement gap between races, if indeed i-genes coding different levels of intelligence are more predominant in one race over another. The pure genetics argument is daunting.

Why Genes Cannot Explain the Achievement Gap

If you are depressed, about to pop the champagne cork, or asking the NAACP to sue me to oblivion--stop. The simple picture I drew above has many holes in it, that upon closer examination, prove the genetics argument to be unsound.

William Dickens of the Brookings Institute has the following to say on the topic:

If researchers were able to identify all the genes that cause individual differences in school readiness, understand the mechanism  by which they affect readiness and the magnitude of those effects, and assess the relative frequency of those genes in the black and white populations, they would know precisely the extent to which genetic differences explain the black-white gap. But only a few genes that influence cognitive ability or other behaviors relevant to school readiness have been tentatively identified, and nothing is known about their frequency in different populations. Nor are such discoveries immanent.

So much for the direct arguments. We don't know any relevant facts about the genetic component of intelligence; therefore if we were to support even the oversimplified genetic argument above, we would have to do so relying on circumstantial and indirect evidence. Traditionally, this evidence has been test scores, including standardized tests and IQ exams. Many have tried arguing backwards from the results of such testin to the genetic root. However, there is ample evidence to suggest that such a backwards supposition would be incorrect.

Dickens writes:

Over the past century, dozens of countries around the world have seen increases in measured cognitive ability over time as large or even larger than the black-white gap... The score gains have been documented even between a large group of fathers and sons taking the same test only decades apart, making it impossible that the gains are due to changes in genes. Clearly environmental changes can cause huge leaps in measured cognitive ability.

The case seems to be shut; genes don't exclusively predict academic achievement. But to say that is only to refute an explanation, not provide one. Genes may exhibit predisposition to cognitive ability, but, given the fact that they do not explain the major differences between the intelligence of two groups, something else must account for the gap between minority and white achievement on exams.

In my next post, I will present Dickens alternate model, which stipulates that "multiplier processes" are available to Whites, and this difference in environment accounts for exaggerated differences in test scoring.

Erik Olsen of the NYT video unit muses in Lens over the effect of music within a film or video production:

A general rule for us — if there is such a thing — is that the music should be appropriate for the story and not overwhelm the story or gratuitously play on people’s emotions. What constitutes gratuitous use? That good question is hard to answer. Music can be effective at creating a particular mood, or, in more extreme cases, it can also signal whether a particular moment should be construed as happy or sad. Purists might argue that the picture and natural sound alone should be used to establish the emotion of a scene, but is that true?

After considering a few examples, Olsen's lackluster conclusion is that

in the end, it is the characters that count: the setting they are in and the challenges they face. The key is to be judicious and practice restraint, to seriously consider when music is appropriate and when it is not. Music used well will enhance a piece without changing it or detracting from it.

He fails to answer the question he starts with and instead offers an overstatement of the maxim "it depends."

Let's rephrase the question, and in doing so, clarify it: Should music be used to establish the emotion of a film or video scene? Olsen ought have seen the answer in his own statement, "it is the characters that count." If it is true that the characters are the significant element, then it must also be true that the characters should establish the emotion of a scene, likely by being sympathetic or antagonistic. In which case, music must not establish the emotion.

Even so, it is also clear that characters are not exclusively fundamental to the emotional pallet. Following Aristotle, the fundamental piece is the plot or the causal sequence which determines where characters of a certain type end up in certain situations. For example, a story of a lovely woman who is faithful to her husband and does nothing risqué may not arouse much emotion--the story will be dull, for the most part. However a story of a lovely woman who is unfaithful and ruins her husband's life--while maintaining a sympathetic characterization throughout--will cause a great stir in emotions. This latter example is of Flaubert's Madame Bovary, a book which poses many difficulties for a reader who wants Aristotle's categories to sit undisturbed, but also tends to prove the point: characters are interesting only insofar as they serve a discreet purpose in the storytelling.

So what do we make now of the question, "should music establish emotion in video?" Obviously the immediate choice whether to overlay music on a scene does depend on what the editor wants to achieve emotionally and how--but that is more a description of what happens in the editing process than a judgement on how one should construct a scene.

Generally speaking, we should instead look at what Olsen is describing. He is describing the process of montage editing. He is describing taking one finished piece of music and combining it with one finished piece of moving photography. The special status of film and video is that multiple artworks can be recombined to construct a new artwork. That process of combination, or montage, is what will ultimately give a finished scene its emotional quality. Therefore, if the emotion comes from the music used, and not the combination of music with imagery, or the combination of various images, or what happens plot wise or in the performance of the actors--if music establishes the mood then we don't have a very good scene on our hands. What we have is a bad or poorly constructed portion of video or film that has been dressed up with music to look better.

Watch the following, first without sound then with, and judge for yourself whether what you see is emotionally effective. (Full disclaimer: I think this clip is bullshit.)

FOXP2 is probably the single most important gene a non-biologist with an interest in language should be interested in. Discovered in 1998, FOXP2 has now been bred into laboratory mice by the Max Plank Institute for Evolutionary Anthropology. (Max Plank, by the way, is under the co-direction of Michael Tomasello, a comparative psychologist whose work on language offers a strong alternative to Chomskian linguistics.)

First some background. The gene is essential in embryonic development; all mammals have it and depend on it for proper formation of the lungs, stomach and brain. As with many genes, there are slight variations between the species. Hence, humans have one version of FOXP2 and mice have a different version. As it happens, a non-fatal mutation in the human version causes specific deficiencies in linguistic ability--a discovery that led biologists to believe that the human gene enables necessary brain developments that allow for speech. So they carefully replaced the version naturally found in mice with the human FOXP2 to see what happens.

The result of this experiment is fascinating, and Nicholas Wade does a superb job in explaining the outcomes:

Despite the mammalian body's dependence on having its two FOXP2 genes work just right, Dr. Wolfgang Enard's team found that the human version of FOXP2 seemed to substitute perfectly for the mouse version in all the mouse's tissues except for the brain.

In a region of the brain called the basal ganglia, known in people to be involved in language, the humanized mice grew nerve cells that had a more complex structure and produced less dopamine...humanized baby mice, when isolated, made whistles that had a slightly lower pitch, among other differences.

That "slightly lower pitch"--though one might be unexcited upon first hearing of it--has enough scientific relevance to show that there are observable differences in the human genome which are by themselves responsible for language.

It's truly mind boggling, especially if you think about it as follows. Imagine we discover all the genes necessary for human speech. Theoretically, we would then be able to trace the activation of those genes to specific tissue development, and from there we could definitively say what structures in the brain are responsible for the production of speech, and how. We could then infer the functional mechanisms of language, and possibly build machines to mimic those functions.

All that is only sci-fi for now; but as of today we know with certainty that FOXP2 is the first of many new bio-linguistic discoveries to come.

Image from: St. Clair, R. et al "The Basal Ganglia and Serial Order of Communicative Signs"