True. That's why I was careful to note that words refer to concepts, not that they are concepts. I have indicated more than once in this thread that definitions must to depend upon words because there never exists a bijective mapping between synonymous lexemes in distinct languages. Even modern languages clearly related (e.g., Germanic languages) and for words with the same etymology, such as god and Gott, there is significant semantic overlap but there is not identity.
The capacity to form concepts without language is extremely limited. Concepts are categories. The concept "computer" is an abstraction that does not refer to any specific computer yet refers to all computers. Perceptual experience cannot allow for the creation of such an abstraction. In general, there is still much debate over what aspects/properties/components of various objects of various types are more central to classification of some instantiation of some object as belonging to category x vs. y. A saw is an abstract concept. There are many different kinds of saws that come in different sizes and are used in different ways (e.g., a chainsaw vs. the saw one would find in a swiss army knife). What is it about saws that make us classify them as saws, yet not do so for objects with similar features (such as serrations on a bread knife or combat dagger/double-edged tactical knife)? Concepts are categories with vague boundaries and usually with more than one prototypical exemplar.
The perceptual experience a baby has when viewing a particular computer is no more salient in the construction of a concept than is the perceptual experience of looking at a TV screen or even a picture. Concepts are not created through primarily through experience but through language. In fact, it is impossible for there to exist anything like the representational range of concepts the human processing system is capable of without language. Perceptual categorization without language requires greater similarity between members in a category and fewer categories. By fewer, I mean that for a cognitive system capable of conceptual processing but not of language categories are either very specific (usually because they are salient to everyday experiences such as "food") or very general (usually because they aren't important). Even more important, the concept of "self" isn't just like any other abstract concept. It allows perceptual experience to be interpreted as consciously subjectively and thus distinguishing between the experience of the perception and the awareness of the perception.
It is extremely difficult for cognitive systems (from computational intelligence programs to canines) to categorize and impossible for any cognitive system without a brain to process concepts. Without language to describe the functions, similarities, and features that make up the various instantiations of "computer" to be conceptualized as "computer" rather than TV, calculator, or even window? What is it that a pre-linguistic understands that allows distinct perceptual experiences to be classified as instances of the concept "computer" yet not others?
Nothing. They can't do it.
All concepts are basic. They are categories we use to categorize and structure are sensorimotor and conscious cognitive processes. The only thing that can make a concept more "basic" is the subjectively salient experiences of similar sensorimotor experiences (like "pain" or "taste").
It is not likely that a pre-linguistic baby can distinguish between features of any given perceptual experience into any clear distinct categories. That would require concepts, and these build up over time as abstractions away and apart from specific experiences. It is essential to the development of concepts that experience is no longer consciously understood as related in any way to the category that instantiations can then be understood as conceptual representatives of.
So are mice.
What is it that we don't know?
