Cider and beer residue on the side of a glass. Photo credit: http://www.flickr.com/photos/livenow/27
At the heart of cognitive science is the notion of a representation: How the mind represents the world during perception and how we relate to these representations when we think.
Representations are the objects of thought, the building blocks of mental experience. Representations have been considered imagistic, propositional or sentential by philosophers going back as far as Aristotle. But only with the advent of cognitivism in the 20th century, has the explanation of mind gone further than mere representation. Cognitivism supposes that representations are aggregates of neural firings and synaptic strengths that also instantiate computational algorithms and symbolic meaning. It is the syntactic and semantic structure of representation that accounts for the systematicity and productivity of thought. This explanation is unapologetically non-reductive, accepting the reality of psychological states such as beliefs and desires as much as the Ca2+ and NA+ ions that underpin them (see Fodor & Pylyshyn's classic article for more details).
As a contrast to cognitivism, connectionism claims that there are no objects of the mind. It supposes that there is no neurological underpinning to beliefs, hopes or desires, because beliefs, hopes and desires don't exist. In fact, connectionists suppose that our talk of mental images, thoughts and so on, whilst pleasurable to discuss, mean nothing in terms of a scientific explanation of the mind. Instead, the only way to understand the mind is to analyze the brain biologically and to formulate our views from discoveries found mechanistically within.
I had assumed that Endel Tulving--the father of episodic memory--was a bog-standard representationalist until reading his article "Coding and representation: Searching for a home in the brain" in The Science of Memory: Concepts. I have believed this because of many instances of representationalist discussion, for example, he has written that the function of episodic memory is to make propositional information available to an organism. Propositional memory has a truth value (i.e. discrete) and can be introspected and considered. Propositional knowledge can be acquired in a single act of perception or thought--contrasted with procedural or skill-based memory that requires repetitious behavioral practice. (Tulving, 1984, 224)
Now, in a small chapter (2007), Tulving says something that makes me question his views.
Tulving considers the nature of the residue left by experience in the mind. Residue is the term he uses by preference, but he acknowledges other frequently used terms such as 'representation', 'coding', 'engram', 'memory image' and 'memory trace' as synonyms. He claims that there is no such thing. Representations, he says, are not physical entities (67).
He defines a memory trace as:
"the neural change that accompanies a mental experience at one time (time 1) whose retention, modified or otherwise, allows the individual later (at time 2) to have mental experiences of the kind that would not have been possible in the absence of the trace. (Tulving, 2007, 66)"
Tulving makes clear that he is referring to "cognitive memory, the kind of memory that has to do with mental experience.... nothing to [do with]... skill learning, conditioning, priming and simple forms of associative learning" (2007, 66). He offers an analogy.
"...think of drawing a straight line. After you have drawn it, the line exists physically with all its properties. Then you grasp the pencil again and make the same line a bit longer. After you have done it, the 'second' line exists physically with all its properties. The difference between the two does not exist anywhere other than in your mind." (2007, 67)
Map of Paris t1
To put Tulving's analogy in context, consider how we learn more about a city. Presumably our mental map of Paris does not change entirely, instead it gets bigger and more comprehensive after we have driven about and seen how more of the roads fit together.
Map of Paris t2
Whilst this explanation might make sense of some learning, there's something problematic with Turing's account of representation. Consider this sentence again: "the difference [between the two lines] does not exist anywhere other than in your mind". How is this expanatory? Understanding how our mind works in a physical body, means figuring out how we can conceive and think about line (t1) and line (t2) in all the complex conceptual and mereological ways we can, yet our brain appears to behave more mechanistically. This is precisely the problem of connectionism as an explanation of the mind. The emergent properties of my thoughts about line (t2) include it's physical properties of length and width, but also abstract properties such as it's relationship to (t1) and all sorts of complex conceptual ideas about identity, numeracy etc... It's difficult to see how to move from Tulving's simplistic memory trace to the subjective capacity of thinking.
What makes cognitivism so compelling is that physical changes in the brain are simultaneously a physical object, a symbolic object and a semantic object--potentially contributing to a many symbolic objects. Surely removing the physicality of representations leaves us either with either mysterious dualism, where representations are individuated by some hitherto unknown mental force or a dispositional account of mind that approaches analytical behaviorism? Tulving states that a "memory trace is something that makes something else possible" (67), what is the something else that is made possible? There is room for counterfactual analysis in mental explanation, but what enables such explanation is the existence of discrete physical entities to which we refer.
In sum, Tulving thinks he's defending an anti-reductionist account of the mind by supposing that representations aren't physical objects. But he offers nothing suitably explanatory in return.
Tulving, E. (1984) Elements of episodic memory. The Behavioral and Brain Sciences. 7. 223-268
Tulving, E. (2007). Coding and representation: Searching for a home in the brain. In H. L. Roediger, Y. Dudai, & S. M. Fitzpatrick (Eds.), Coding and Representation: Searching for a Home in the Brain (pp. 65-68). New York, NY: Oxford University Press.