A Research Proposal for The Intimate Study
Charles Peirce argued that research in physics is of three kinds: first, the making out of new phenomena; second, the investigation of their laws; third, the measurement of constants. If the science of learning is to be like physics at all, should it not be so in permitting this latitude among the legitimate aims of research?
Peirce goes further in valuing as most important the making out of new phenomena. An example from current controversies will make clear why I find his judgment congenial. Though students of the mind from one stance or another may quibble over Piaget’s formulations of development, none will deny him foundational status in the modern study of mind precisely because he discovered the regularities of ignorance. [How ironic that, much as Einstein’s ‘invariance theory became ‘relativity,’ Piaget is popularly known for discovering ‘conservation.’]
I part company from Peirce when he avers that the discovery of new phenomena is largely a matter of chance and when he makes that most dangerous of speculations, ‘the new phenomena which now remain to be discovered are probably only of secondary importance.’ My view comes from a contemporary vision of doing science which in part Peirce inspires, but in which he could only partly share and a vision wherein the critical issue is the relationship between data and theories.
Data and Theories
We are fortunate to have lived in a period of rapid scientific change, one consequence of which is that process itself has come under scrutiny and been rendered more nearly comprehensible. Our climate of opinion now takes for granted that essentially sociological perspective on scientific change advanced by Kuhn. And if such micro-cultural relativism leaves one uneasy, as it does me, there is still the profound ideological antidote of Peirce’s vision of truth as the convergence of opinion over time in the community of minds. Gould [see “The Continental Drift Affair,” Natural history February 1977] has described the recent paradigm shift ingeology (i.e. to the general acceptance of continental drift as a consequence of the plate tectonics theory of geodynamics) as a pellucid example of scientific change. The essential points of his description reveal the interrelation of data and theories in the most important scientific work.
Cast your mind back 20 years. Recall those popular books by Velikovsky, Earth in Upheaval, and so forth, wherein he argued astrophysical explanations for data no other theories could explain (e.g. peculiar distributions of African-like fossils in Scandinavia, etc.). He argued unconvincingly that astrophysical events had triggered continental migrations because there was no conceivable mechanism on earth to explain how continents of granite could plow through the seas of basalt underlying the great world oceans. He was ridiculed and the data to which he called attention were dismissed.
Gould recounts how a few mere years ago a new theory of geodynamics, proposing a mechanism which was capable of explaining the gradual separation of continents, raised again to prominence these data Velikovsky had invoked. The new theory swept the field of geology so that it is now a different science from what it was. The reason was not Velikovsky’s data. The coerciveness of the plate tectonics theory derives from a prediction of the relevance of a new class of data (the directional magnetization of patterns of deep sea core samples), the discovery of this new phenomenon, and the lack of any conceivable mechanism other than sea-floor spreading for interpreting the results found. The pattern of change is this: a new theory is proposed which demands examination of new data; those new data exhibit phenomena which are interpretable only in terms of that theory; the new theory, in rendering one of several prior theories a consequence, selects between what had been previously competing interpretations of older data.
The computational theory of the self constructing mind will play a role in the science of learning analogous to that of plate tectonics in geology. I expect to find in the class of data I am now collecting sequences of events which be interpretable as the interaction of concrete representations of ideas with specific events in terms of the progressive elaboration of those concrete ideas (debugging to meet new environmental requirements) and insight through analogy (abstraction of procedural control structure through substitution of terminal values and lowest level procedures). I believe these data will not be interpretable any other way.
Such data as I now gather may be vulnerable to criticism because they are of my daughter. One might note that geologists studied sea-floor spreading in the mid-Atlantic, not in their own back yards. Why do such a study as mine is ? I would ask a similar question: why did geologists not first study sea-floor spreading on the moon ? This lame joke is meant to suggest the important issue is not interpersonal involvement but the even more primitive question of what is the whole phenomenon to be explained and under what conditions is it possible (or impossible) to collect those data which are relevant to explaining that phenomenon. The geodynamicist must explain one mechanical system, the earth. There is only one earth and that is relatively convenient to observe. I believe one mind is a complete system to be explained. If it is necessary in tracking concrete development to observe that mind in all its settings and situations, it is not a mere convenience but a requirement that the mind be available to observation nearly all the time. For a researcher to win parent’s confidence and gain permission for such a pervasive study would be most difficult and most likely impossible. I conclude it is a practical necessity that such a study be of one’s own child.
If one argues further that the geologist takes core samples near the mid-Atlantic ridge, not everywhere, I admit freely that the analogy breaks down. The selective principles in my observation and data recording focus first on the locating the actions, objects, and situations to which may be plausibly ascribed specific ideas in Miriam’s mind; second, on observing the way Miriam’s ideas change at a low level of detail to permit plausible inferences about the immediate or proximate causes of these changes.
In conclusion these few words: men have long argued over the relative import of intrinsic and extrinsic factors in development. With the newly specific ideas of artificial intelligence, I hope to trace the interaction of psyche and settings to see how of the concrete world we encounter:
"Nothing of him that doth fade But doth suffer a sea-change Into something rich and strange" the human mind.
- Written in 1976, as part of a proposal for that doctoral research completed as “One Child’s Learning” in 1979. Unpublished.
- Subsumed in chapter 2 of “Case Study and Computing.” Ablex, 1996.
- the final lines are from Ariel’s song in The Tempest, by Shakespeare.
When Jane Goodall was brave enough to study chimpanzees in the wilds of Gombe, it seemed only modestly adventurous to go outside the lab and study children in the wilds of my own backyard.