as is patent in the development of talk of dictionaries and concept-stores in the brain. It is self-defeating to appeal to pragmatism, as if experiments could validate analogies independently of laying down verifiable criteria of identity and difference (the extension of hydrodynamic analogies to electricity theory is not self-validating or self-confirming ).
In this book we sketch an overview of some of the most widespread and damaging conceptual confusions in cognitive neuroscience. Almost all the chapters of the first edition have undergone revision and addition, sometimes minor, sometimes major. Some revisions have involved correction of previous error (e.g. phantom pain location), others involve additional explanation and elucidation. Most new sections involve reference to and critical discussion of papers published on the subjects of concern over the last two decades. We provide a handbook that we hope will be of service for neuroscientific research. Above all, we try to exemplify an array of analytic methods of conceptual clarification. These methods require no laboratory space and need no grants. All they require is an open mind, a willingness to raise a question for every answer, and patient intellectual effort. This can save a young scientist from committing his or her life to an enterprise that was doomed to failure before it even started (e.g. the quest for the neural correlates for qualia – see chapter 11 – or the search for the locus of the self – chapter 14).
We have added to this second edition numerous figures, charts and tree-diagrams, comparative lists and tables. They are meant to facilitate obtaining an overview of a conceptual domain – providing a synoptic presentation that can be taken in at a glance. They are often no more than illustrations to the argument of the text, not substitutes for it. They sometimes oversimplify for purposes of surveyability. So they should be used with care. With this proviso, we believe that they fulfil a useful function in aiding understanding.
M. R. Bennett
P. M. S. Hacker
Part I Philosophical Problems in Neuroscience: Their Historical and Conceptual Roots
Preliminaries to Part I
1 Philosophical Problems in Neuroscience: Their Historical Roots
The more than two thousand years of the history of what became known as cognitive neuroscience shows that the discipline is grounded in correlations of human behaviour, reflecting psychological powers such as seeing, hearing, remembering, etc, with areas of the brain and their function. Such correlations are often erroneously taken to show that psychological attributes pertain to parts of the brain (the occipital cortex perceives, the hippocampus remembers), whereas the correlations are with the behaving human being, who perceives and remembers.
Galen
In the second century ad Galen identified the proper functioning of the enkephalon which is necessary for the mental powers of humans. This was based on careful clinical observations of the many injured charioteers and gladiators he had access to, followed by careful pathological examination of their bodies. Galen may be thought of as the initiator of what became known as the clinico-pathological procedure in neurology, whereby correlations are sought between abnormal behaviour and a diseased state of the body. In this case Galen observed that the ability to think and reason were impaired when damage had occurred to the brain, but not to other parts of the body. Unfortunately, when he passed on to the question of what particular parts of the brain were necessary for reasoning he departed from the clinical-pathological method, and guessed that it was the ventricles, considering the cortex as merely the wall of the ventricles.
Nemesius
In the fourth century ad Nemesius believed he could affirm and extend Galen’ s claim that the ventricles were correlated with possession of human powers using the clinico-pathological approach. He claimed to show that injury or disease to each of the ventricles led to behavioural deficits indicative of the failure of a particular power. For example, if the front ventricles are harmed then the senses are abnormal but thought remains normal. This example of incorrect conclusions drawn through application of the clinic-pathological technique indicates how difficult the search for correlations can be, and is manifest even in the search for the primary visual cortex, for which a consensus was reached only towards the end of the nineteenth century.
Thomas Willis
In the seventeenth century Thomas Willis, a master of the clinico-pathological technique, showed that it is the cortex of the brain, not the ventricles, that is correlated with the possession and exercise of psychological powers. Willis often studied the behaviour of his patients over many years, writing up detailed reports of abnormal behaviour indicative of the failure of a psychological power or a group of powers, and then correlating these with abnormalities of the cortex revealed on dissection at autopsy, all of which he published. Perhaps the most spectacular of these was his identification of congenital mental disabilities with clear cortical abnormalities (Figure 1). But even more specific behavioural maladies, such as long-standing one-sided paralysis, were correlated at autopsy, in this case with unilateral degeneration of the cerebral peduncle. It was the quality of the careful observations that Willis made which allowed him to identify cortical lesions as opposed to those organs such as the spleen and lungs that were claimed by others to be the basis of disorders of behaviour. By the time Willis had completed his clinico-pathological correlations, together with detailed anatomical drawings, there was no doubt that a normal cortex was necessary for normal behaviour. But although the ventricles had been displaced as centres of interest in this regard, the question of which part(s) of a normal functioning cortex are correlated with the behavioural manifestations of particular psychological powers had not been elucidated.
Figure 1 Illustrations from Cerebri Anatome of Thomas Willis, showing on the left the underside of a normal human brain and on the right of a brain from a person suffering from congenital mental weakness. Reproduced with the permission of the library of St John’s College, Oxford
Clinico-pathological correlations in the nineteenth century
The nineteenth century saw a flowering of the clinico-pathological correlation of cortical areas with psychological powers, such as vision and speech. It might seem straightforward to identify abnormal behaviour(s) uniquely due to the loss of a psychological power, but this is not necessarily so, as behaviour exhibits such a variety of different forms.1 Furthermore, identifying at an autopsy the unique site of a lesion or degeneration is difficult at any time, let alone with no more than the tools available in the nineteenth century. Even when considering a comparatively straightforward power in this regard, such as vision, it took fifty years before the primary visual cortex was agreed upon as located in the occipital lobe.2 Nevertheless, significant progress was made by Broca, as well as Fritsch, Hitzig and others, together with Sherrington’ s researches on primates.
Single neuron investigations and fMRI in the twentieth century:primacy of behaviour
In the twentieth century the correlations revealed following careful application of the clinico-pathological technique were followed by neuroscientific investigations into the neural mechanisms operating within the identified cortical areas. The introduction of new techniques by Adrian and Eccles in the first half of the twentieth century allowed for a new level of spatial resolution relating psychological powers and cortical function, namely that of a neuron within the already identified
