Chapter 5 & 7 - Adam & Angela

Chapter 5: Emergent properties and Connectionism

The connectionist model is an alternative to cognitivism that places the emphasis of cognitive function on the emergent activity of neurons. Connectionist models offer advantages over 'top-down' cognitivist models in that they are robust, parallel, distributed, and offer high potential for learning. They are also closer to biology.

There is a difficulty in trying to assign discrete symbolic states to a connectivist system, since its properties are dependent on the distributed state of the whole system. In connectionist models, symbolic computation is replaced by whole-system states described by numerical operations: patterns replace symbols, and we move from the symbolic to the sub-symbolic as the main level of interest in neurological activity.

There is nonetheless an exploration of how symbols and sub-symbolic models can be related. Approaches such as complementary 'top-down' and 'bottom-up' views are suggested, as is the use of symbols as a method to describe different states of the whole connected neuron system, or its sub-parts. Cognitivists argue that a symbolic model thus remains valid for describing higher-order operations, and complex abstract systems such as language. The counter argument to this is that we still define cognition in this case as logical reasoning, which is too limiting given the multiple cognitive processes at work in the formation of mind, as well as excluding other distributed systems that display cognitive properties (the example of immune systems is given).

Emergence is also used to address the question of sequentiality in the five aggregates of Buddhist thought. Rather than a sequential process of perception and reaction, we can see the aggregates as a description of the emergent state of the mind at a given moment, a view more closely aligned with our understanding of brain functions, as well as our lived experience.

Chapter 7: The Cartesian Anxiety

This chapter questions the notion of representation in cognitive science. We are accustomed to conceiving of the world as pre-given, with properties which can be 'recovered' and internally represented by a cognizing agent. The authors distinguish between a 'weak' form of representation, or 'construal', in which something is said to be 'about' something else without considering how that semantic meaning is attained. The more problematic form is 'strong' representation, or the generalizing of the weak form in order to develop theories of language or cognition. This form is problematic because it makes the above assumptions of a pregiven world with properties that can be represented.

Cognitive Realism seems to present a solution to the Western philosophical opposition between realism and idealism, but still retains the problem of the 'image of mind as a mirror of nature'. This is problematic in that it treats information as a property of the world to be processed by the mind: the view of mind as an information-processing device with input and output. But, as noted by Minsky "it makes no sense to speak of brains as thought they manufacture thoughts the way factories make cars. The difference is that brains use processes that change themselves - and this means we cannot separate such processes from the products they produce". In other words, the world as perceived by a cognizing agent must be considered to be produced or at least influenced by those cognizing processes.

We are now seemingly left not only without a solid inner platform from which to relate to the world (the unified self), but also without a solid outer ground to relate to. This is described as the Cartesian anxiety, since the problem only arises when we conceive of a solid outer world which we are barred from knowing, an inheritance of the Cartesian notions of clearly defined inner self and outer world. Once again, this features in Buddhist philosophy, as they describe a grasping of a solid world as well as the grasping for self. By letting go of the need for a solid ground to precede from, we are open to lived experience.