Here’s how it begins..
Rules and Similarity in the Development of Category Learning in Children and Infants
I give you now Professor Twist,
A conscientious scientist,
Trustees exclaimed, “He never bungles!”
And sent him off to distant jungles.
Camped on a tropic riverside,
One day he missed his loving bride.
She had, the guide informed him later,
Been eaten by an alligator.
Professor Twist could not but smile.
“You mean,” he said, “a crocodile.”
Ogden Nash, I’m a Stranger Here Myself (1938)
Ogden Nash never informs us what his conscientious scientist was sent into the jungle to do. But we may guess, based on our familiarity with other jungle-bound scientists, that Professor Twist was probably a biologist, sent there to classify the fauna he found. Classification can be done in two ways. One may place like with like, as when we see the similarity between a crocodile and an alligator. Or else one may follow rules, as when a biological taxonomist draws up the list of the defining characteristics of Crocodylus niloticus and Alligator mississippiensis. It might seem obvious that these are two very different processes; The first is fast, flexible and automatic while the second is precise, focused and deliberate. But how do they work, how are they related and how do we determine which type of categorisation to use and when?
Classification by means of similarity seems like child’s play, but this belies its power. It is the work of an instant to identify to an alligator, which is just as well because it might be a matter of life or death. We could still recognise a crocodile if it was bright pink, or half an inch long, a character in a cartoon or even if it had been turned into a handbag. Even young children can to do this and all without being able to deﬁne or articulate what is the essence of a crocodile, or even needing to. Similarity allows one to generalise. If you had only ever met alligators you would have no difficulty in recognising that crocodiles were their cousins and if you were told that alligators have integumentary sense organs, you would be likely to conclude that crocodiles did so too. One may mistake a crocodile for an alligator but one would not make the mistake of thinking that either was friendly (Actually although crocodiles are foul-tempered, alligators rarely attack people unless sorely provoked.). Although sometimes, like perhaps the unfortunate Mrs. Twist, one may be fatally misled by their shared similarity to floating tree trunks.
And what of classification by rules? Well, take for example, alligators and crocodiles; The similarity of the two is such that it takes a rule to tell them apart. There are a few rules you could choose from; Alligators have U-shaped snouts while crocodiles have pointy V-shaped noses. Crocodiles have teeth in their lower jaws that protrude up the sides of their mouths while alligators do not. Or, since Professor Twist did not see the smiling face of the beast that ate his bride, we may reason that he reasoned by another rule; ”If one is in Africa then they are crocodiles.” All involve the application of abstract knowledge to a restricted set of the available information. The ﬁrst two involve focusing on a small number of perceptual features and the third uses a verbal rule that concentrates entirely on context. While classification rules may then appear the remit of quintessentially rational adults, they are not purely for pedants. By building categories that are based on single simple features, or deﬁnitions, we acquire manageable, manipulable blocks with which to reason. Rules can be chained, nested, combined or even contradicted to build up complex knowledge about the world. By these means, conceptual thought is extended beyond its perceptual foundations in a way that is distinctively human.
Like Professor Twist, the present author is also setting out conscientiously on a mission of scientific classification. In fact the aim of this thesis is a classification of classifications. It sets out to tell the difference between rules and similarities in category learning, to understand how each process works and how they are related. A biologist seeking to understand alligators and crocodiles can investigate their anatomy, physiology, life-cycle and ecology. He or she can also appeal to the theory of evolution and look back into the fossil record. Likewise, the cognitive scientist can meticulously investigate fully realised adult abilities but may also appeal to comprehensive theories of these abilities and look at the early origins of adult-like performance in the abilities of children and infants. This thesis adopts that developmental approach. The rest of this chapter provides an introduction to the relevant theoretical and empirical literature and motivates the particular developmental approach taken. The following four chapters each presents a self-contained piece of empirical research that aims to address some of the questions raised. While the ﬁnal chapter summarizes the results that are found and considers what more general conclusions can be drawn. The rest of this section lays out in more detail the particular the aims and objectives of this research and provides a
brief outline of the literature review that follows.
But I won’t trouble you with the following 300 pages because it becomes substantially less interesting.