Jerome A.Feldman
From Molecule to Metaphor
A Neural Theory of Language
Mit Press 2008

Feldman 15-38
The information processing perspective: Neuroscientists speak of neurons as processing information and communicating by sending and receiving signals. They also talk of neurons as performing computations. In fact, neural computation has become the standard way of thinking about how the brain works. But neurons are just cells, tiny biological entities that are alive and function by means of chemistry. Why can he say that neurons process information and perform computations?

Since neural computation and the information processing perspective on the brain centrals to a neural theory of language, it is important that I say, in as simple, clear, and straightforward terms as possible, just what I, and others who study the brain, mean by neural computation.

Amoebas as information processors: three ways of thinking about an amoeba:
• A chemical factory: an organised system of about a million complex molecules. This view is required for many scientific purposes, but additional perspectives are needed for a full understanding.
• A creature with needs, desires, and goals: the chemical perspective in itself doesn't help much when we are trying to understand the amoeba's behaviour. A common way of modelling behaviour is through a personification metaphor: try thinking of an amoeba as if it were a person with needs, desires, goals, and ways of satisfying them... From such an empathetic perspective, one might say that the amoeba is trying to find food and avoids danger. To understand how the amoeba behaves, both the chemical and the empathetic perspectives are required at once.
• The information processing perspective: now let's think about the amoeba from an information processing perspective. This is another generally useful way of talking about living things, and how they satisfy their needs and goals. What information does the amoeba use to survive? Do amoebas communicate with one another? As observers of amoebas, we may decide to ask such questions, and doing this requires an information processing perspective.

From this perspective, the two types of chemical reactions (to food and to irritants) can be seen as enabling the amoeba to distinguish two kinds of inputs. In general, we always try to understand new things by relating them to a familiar concepts. The information processing stance is often useful because we have a rich knowledge about computing, memory, and learning, and we could use this wisdom to help us understand what amoebas and other living things do. An additional set of information processing questions concerns communication. Since amoebas do not reproduce sexually, they normally have nothing to communicate about, but other single celled creatures do communicate using molecular signals.

19 Communication and co-ordinated evolution
Communication between cells is a major revolutionary advance and a prerequisite for the appearance of multicelled creatures like ourselves. Individual cells survive by carefully controlling their internal chemistry and it goes against their nature to allow outside agitators. Of the 4 billion years since life began, about two thirds was required to evolve the simplest multicellular organism and their coordination mechanisms. The basic mechanism of the communication is molecular matching…

...the emission and subsequent recognition of a signal molecule is the simplest form of communication among living things.

20
But communication requires more than just having cells recognize things in their environment that are either good or bad for it. Communication occurs only when the sender and receiver agree on the nature of messages.
This inherently requires co-ordinated evolution.

We can often ignore considerations of the physical details and study communication from an information processing perspective, which specifies what counts as an input signal, output signal, recognition, reaction, memory, learning, communication, and so on. The information processing perspective is crucial... as we look at yet another kind of cell: the neuron.
The idea of cells communicating information using small molecules as signals is also the necessary first step in making sense of neural computation.
But before we turn to neural information processing, we need to understand more about information processing.

21
…one crucial point here is that the information processing perspective is always metaphorically imposed - it is one way of understanding a complex system in terms of something completely well-defined and very well understood. There is no abstract information in the neurons themselves; like amoebas, neurons are cells that work through chemistry.

An information processing metaphor creates an abstraction, in the sense that it abstracts away from - that is, ignores - any non-informational content of the physical system being studied, whether animate (like neurons) or electronic (like a computer). Any living system carries out specific physical processes in which distinctions are made towards satisfying needs and goals, and so it can be studied from the information processing perspective.

Information processing metaphors can be extremely powerful tools for understanding physical systems… However, we can't understand the real animals or any other complex system using only the computational stance. (example of the difference between the information processing abstraction and physical reality and how both are needed to understand what is going on: the telephone system…)

22 Abstract information processing models – Turing machine
29 Computational models
30 Simulation
33 Functionalism: No one believes that the computer simulation of a weather pattern is itself anything like the weather itself: the ever-changing clouds, the heat, the rain in your face, the oppressive humidity, and so on. The model is clearly distinct from the reality being modelled.
But when we use a computer programme to simulate some function of the brain, we get into some delicate philosophical questions about exactly what is being done. One possibility is that, like the weather simulation, the computer is simply being used to carry out a formal, computational description of some process of the mind. The process is understood as being carried out by the physical brain, which is quite different from the programme used to model it.

A second possibility, however, has become a major intellectual position within Anglo-American philosophy, generative linguistics, cognitive psychology, and artificial intelligence. This position is called functionalism. In its strong form, it claims that the way the mind is physically embodied in the brain is irrelevant to the study of mind. Functionalism as principal is the opposite of an embodied theory, which suggests that everything important about language depends on the brain and body…

As we saw in previous chapters, scientists are all this study nature using various perspectives, and a functional analysis is usually involved. Almost everyone agrees that a functional level of description is needed for language and thought. Philosophical functionalism holds that everything important about language and thought can be understood completely using information processing models, without looking at the brain at all. An even stronger position claims that any information processing system of sufficient complexity will automatically have all the mental powers of the mind, including consciousness. This stance is also called strong artificial intelligence.

37 neural information processing: Neural information processing systems are sufficiently different from their electronic counterparts that it has proved necessary to develop special theories and simulation techniques for the neural case… neurons are a million times slower than electronic components. But each neuron is connected to thousands of others, most of which are active most of the time. In contrast, electronic computers are extremely fast, but have only local effects and only a tiny fraction of the elements are simultaneously active.

This difference in basic computational character has the most profound consequences for project of modelling thought and language. Because the brain is richly connected and profusely activated, there is no such thing as an isolated or purely abstract thought. One idea automatically activates others. In addition, any input of language of perception is understood against a background of ongoing activity and is all this contextual. This has long been understood informally and essential to modern psychological theories of memory and language processing. The main consequence of these findings will be that mental structure parallels active neural structure - connected concepts are neurally connected.

Another crucial property of our brain is its intimate link to our body. Digital computers are designed to computer general functions; brains evolved to control animal bodies. The link between sensation and action remains the dominant function of the human brain. Language and thought refined means of connecting inputs to desired outputs and work through computational mechanisms based on the embodied brain.

Neural systems also differ from electronic systems in that there is no separate programme. Instructions as such do not exist, and control appears as pattern of activation in the network itself. This also has profound implications for learning. Neural systems appeared to acquire knowledge in two ways, weight change (change at the synapses) and structural recruitment (the strengthening of a previously latent connection between active neural clusters).

Feldman
Cognitive Linguistics