Sunday, March 16, 2008

Some arguments that fall flat and why (Pseudolinguistics and Ideal Languages, Part 1)

This is a prequel to a set of articles I have planned on pseudolinguistic claims made by amateur historical linguists. Most of these claims propose that some favored group either spoke the first human language, made some voyage or discovery earlier than supposed in mainstream linguistics, or played a pivotal but unrecognized role in the course of history. Before I begin, I want to draw attention to two archetypal forms of arguments that a purportedly evidence-based but specious argument can take. These will play a role in the discussion and might be immediately recognized by those who have watched any debate on pseudoscientific or supernatural claims.

The first type is often a combination of argument from ignorance and false dichotomy, essentially the argument that some rival hypothesis is currently incomplete. It is often coupled with an insistence that the rival hypothesis can never be complete.

  • Hypothesis A and Hypothesis B are potential descriptions of Phenomenon Z.

  • Hypothesis A does not explain some aspect of Phenomenon Z.

  • Hypothesis B states that Hypothesis A will never explain some aspects of Phenomenon Z.

  • Therefore, Hypothesis B is correct.
There are a few obvious problems with this argument.

  • The argument ignores Hypotheses C, D, E, . . .

  • The truth of Hypothesis B is not predicated on its ability to predict the same aspects of Phenomenon Z.

  • Evidence may yet be found that allows Hypothesis A to explain that aspect of Phenomenon Z.
The failure of this argument was recently brought to light in the debate between Myers and Simmons. Simmons argued that the hypothesis that species arise from descent with modification could not adequately explain the presence of order cetacea (dolphins, whales). At one point in time, this statement may have had some argumentative weight: marine mammals were clearly an aspect of the geographic spread of species that any hypothesis needed to explain and early evolutionary accounts were plausible but lacking in true evidence. However, the evolutionary hypothesis made predictions concerning the genetics, morphology, and geo-historical distribution of marine mammals and their ancestors throughout the fossil record. In time, these predictions were found to be true: as pointed out in the debate, we can now trace some important changes in the respiratory system of marine animals and their ancestors, among other aspects of their morphological history.

The supernaturalist's argument, on the other hand, does not predict
  • The morphological similarities and differences between marine and terrestrial mammals.

  • The distribution of mammalian fossils with marine and intermediary stage features.
Intelligent Design and other formulations of creationism offer no real explanations except to say that what-you-see-is-what-you-get: any distribution of species and features is an equally valid outcome. Its "success" rests entirely upon whether other hypotheses fail.

You can witness this argument in the current blog-debate between Steven Novella and Michael Egnor on materialist and dualist explanations for neurological phenomena. Egnor writes

If dualism is true and the mind is partly the product of the material function of the brain and partly the product of something else, then:

1) There will be some mental phenomena without brain function
2) As brain function is altered, the mind will not necessarily be altered
3) If the brain is damaged, then mental function will not necessarily be damaged
4) Brain development will not necessarily correlate with mental development.
5) We will not always be able to correlate brain activity with mental activity – no matter how we choose to look at it

The discerning will notice that Egnor's hypothesis does not predict which mental phenomena will not correlate with brain functions, which mental phenomena will not be altered by brain damage, etc., and instead its success depends almost entirely upon there being no Hypothesis C for the unnamed phenomena: e.g., that mental phenomena without brain function could be generated by other physical means. According to the current evidence, the materialist hypothesis doesn't even require that dramatic of a revision: so far there has been no evidence of any mental phenomena thats exist without neural function. In the post linked above, Novella does a particularly good job at debunking the one piece of evidence Egnor claims. Novella also points out one not so evident problem with Egnor's hypothesis: determining a lack of brain function rests entirely on the accuracy of contemporary methods for detecting brain activity. Like creationism lurking in gaps in the fossil record, Egnor's brand of dualism continually lurks just beyond of the resolution of brain imaging techniques.

Of course, there could be creationist and dualist hypotheses that do not resemble arguments from ignorance. I'm not familiar with any worth mentioning (e.g., that anyone familiar with the actual evidence would still think worth lengthy discussion), but I don't doubt some are out there. I also recognize Egnor may not be regarded as the most distinguished proponent of dualism: I tried to be careful in associating his characterization of dualism to him. Likewise, Simmons' beliefs about marine mammal evolution have been attributed to him, not all cdesignproponentsists.

There are even some hypotheses that have been accepted by at least portions of the early or modern scientific community that might be formulated this way. In linguistics, Noam Chomsky has famously gone on record as stating that he did not believe evolution possessed the means to generate a language faculty as developed as that of human beings, though he has not proposed an adequate counter-explanation (to my knowledge). His opinion did not stop biologists from looking, of course, and they have spent the last few decades uncovering bits and pieces about animal cognition that lays the foundation for alternative evolutionary hypotheses. A contemporary course that"taught the controversy" on the origin of the language faculty wouldn't likely include Chomsky's premature claim that, in essence, human language is irreducably complex. (Although I'm not sure how much weight it carries, Chomsky may have updated his position to say that language evolved from a reasoning module. There has been ongoing discussion on this over at Babel's Dawn, including this post.)

The second type of argument I want to highlight is more difficult to characterize. It essentially rests on weak or false correlations and other cherry-picked or "salad bar" assemblages of evidence. Although I see the first type of argument most often being used in support of supernatural claims, I must admit this second type is not at all unheard of in mainstream academic discussions.

In fact, the actual hypothesis of the argument is often a legitimate scientific hypothesis. It might be motivated by prescientific, ideological, or supernatural beliefs, but the question asked and solution proposed are typically within the realm of scientific investigation. By focusing on larger claims, the arguments often overlook long-accepted and repeatable measures and thus often require one to reject well-established and independent theories. Paradigm shifts do happen in science, but most often they do not require one to reject data, only hypotheses that did not work and idealizations that were inaccurate (and in many historical cases, long known to be inadequate by specialists).

(Side note: For a overly flashy but nonetheless interesting look at paradigm shifts in physics, take a look at this documentary on the atom: part 1 here. Note that although our (empirically motivated) conceptions of the atom changed, each shift in our understanding was accompanied by a new set of empirical oddities that the new theory did not address. Far from ignoring those problems, much work following the shift tried to tackle them.)

Well known pseudoscientific and pseudohistorical examples of this type of argument can be found in the creationist attempts to account for global flood and the apparent age of the earth. As always, Talk Origins has some excellent pages devoted to these topics: see The Age of the Earth, Radiometric Dating, and Problems with a Global Flood. Because the mainstream account of the age of the earth, the formation of geological features like the Grand Canyon, and the distribution of sedimentary layers rests on multiple independent lines of inquiry and well-known physical properties of matter, proposing a catastrophic or recent origin for any of these things means throwing out a great deal of other evidence upon which much of modern physics, chemistry, and biology rests. Proponents of a young Earth or global flood often attempt to defend their hypotheses from these problems one by one, but what their basic hypotheses ultimately require is a complete rewriting of many modern sciences at the most basic level and it isn't often that one sees these proponents suggesting adequate and functioning frameworks to replace the old. (And considering the amount of evidence in support of the mainstream views, one cannot but forgive scientists in these fields from continuing their work as if there is no evidence at all of a geologically recent global flood.)

Another recently popular example is the 1421 hypothesis, which posits that Chinese explorers visited and mapped large parts of the world (including the Americas) independent of Europeans. Some of the maps in support of the hypothesis certainly demonstrate that Chinese cartographers had some information on Africa and Europe, but whether this information came from discovery by Chinese junks and not from shared contact isn't satisfactorily proven. The lack of precision in map making and writing of the era make analyses of books and maps open to subjective interpretation (e.g., in claiming a Chinese drawing resembles an armadillo). Accepting some of the evidence also means rejecting many established methods of dating and ascribing creators to archaeological sites (e.g., Newport Tower) and accepting claims on uncategorized or even uncategorizable material wholesale without supporting evidence (e.g., isolated undateable poettery fragments). Whenever possible, historians prefer to work with materials and accounts whose usefulness for supported or eliminating theories can be systematically and objectively demonstrated.

It is of course entirely possible that Chinese explorers visited the Americas, and such a hypothesis should be readily testable, but in the end, unlike the Europeans, there's no large patterns evidence to satisfactorily prove that they had been there and returned (e.g., no objectively Chinese settlements in the Americas, no evidence that trade was established and widespread, etc). Its this need for an identifiable pattern to the evidence that keeps these speculations from reaching the mainstream: the 1421 hypothesis doesn't tell us where we should look next. Proponents of this second type of argument typically grasp for as-yet unexplained phenomena and - instead of accepting that they might be unexplainable - stitch them together with weak threads to support their hypothesis. The unsystematic grab for evidence and loose explanations to tie everything together perhaps rightly remind one of conspiracy theorists for whom all actions by the alleged conspirators count as evidence for their theory. The sloppy nature of the argument (hand-picking evidence as it suits the researcher) often leads to false claims (for the 1421 hypothesis, that small pox spread in the Americas before Columbus), but to me, this tendency seems to be more a result of the type of thinking involved than a necessary requirement of the archetype.

The next post in this series will be on methodology in historical linguistics and a few accounts that mainstream linguistics reject. Hopefully, this introduction will help show how these pseudolinguistic claims are related to the pseudosciences and pseudohistories that readers are undoubtedly more familiar with.

No comments: