The Mundane Matter of the Mental Language
J. Christopher Maloney
Cambridge Studies in Philosophy
Cambridge: University Press, 1989 (Paperback re-issue, 2007)
274 p. £35.00. ISBN: 0-521-03929-0
Reviewed by Laure Gardelle
École Normale Supérieure de Lyon
This book is a philosopher’s inquiry into the language in which we think, a topic which has a long tradition in philosophical studies, especially among cognitivists. By thinking, the author does not mean articulated arguments, but the most basic computations and inferences that stem from our beliefs and desires and govern our actions. Thinking, which is defined as a process designed to solve problems, is not performed through overt languages (“actual” languages, such as English), since babies or intelligent animals are able to think and yet do not master such languages. It is probably not innate either, as it is linked to cultural practices. Rather, although it is not acquired by learning, it is picked up gradually.
The author takes a stance against functionalism, which focuses solely on the functional properties of mental states. Following previous studies, he suggests that the physical properties of mental states must be taken into account as well: it is probably through sensation that mental expressions are originally encoded and derive their content. The author defends more specifically the Sententialist tradition. It postulates a language of thought, which he calls Mentalese. This language is inaccessible to consciousness and, like all languages and systems of logic (whether in mathematics or computer sciences), is a symbolic system. It is endowed with a grammar and semantics which generate sentences – not the actual sentences we produce in overt languages, but sentences in Mentalese, of which our thinking consists.
The book is divided into seven chapters. Chapters 1 to 4 are a theoretical introduction to Sententialism, while chapters 5 to 7 are meant to exploit that theory. The author indicates that readers who already have a fixed opinion on the language of thought may start the book with these last three chapters. Chapter 1 argues for a representational theory of the mind. It establishes the traditional difference between our mental constructs and the actual world in order to define mental states as representational structures. Focusing on beliefs because they are a cornerstone of the philosophical tradition, the author shows that behaviour is generated through internal inference, and that perception draws on sentential structures to advance its inferential process. Chapter 2 shows that the number of beliefs in the mind is high but not infinite, and that they are organised in networks of frames and scripts. Following Schank and Abelson, the author proposes that the mind contains topic-specific structures of representation designed to facilitate processing, so that alteration of one belief immediately affects other related beliefs. Chapter 3 focuses on rationality in behaviour, another argument in favour of Sententialism. Chapter 4 specifies a definition of Mentalese in the face of objections raised by previous studies and determines some of its properties through a comparison with artificial intelligence. Chapter 5 expands on the specificities of the mind compared to artificial intelligence systems. Mentalese is shown to be close to those systems, in that the mind is computational, but only the mental language of the mind is truly intelligent, i.e. conveys understanding and feeling.
Chapter 6 explores the nature of Mentalese proper: how can Mentalese sentences mean what they do, and what are the terms that compose them made of? In order to answer these questions, the author first turns to a definition of predication and attribution. Predication in overt languages is a combination of two operations: attribution, defined as the selection of a property for assignment to an object, and the location of that property within a conceptual scheme of related properties. Attribution, therefore, is the semantic counterpart of demonstrative reference: it consists in assigning a property to an object without also conveying how that property is related to other properties. From this, the concept of “sensuous predication” can be posited: it is defined as mere attribution, i.e. as the representation of a property merely as a distinct property.
Attribution and predication are key concepts for the description of Mentalese sentences, for instance to describe what happens when a cognitive agent senses a red apple. The apple’s being red causes a sensory state S to occur in the agent, and the agent sensuously represents “this (i.e. the apple) is that (i.e. red)”. This mental state is a Mentalese sentence, in which the subject is the equivalent of a demonstrative pronoun and the predicate is pristinely attributive – in other words, sensuous predication identifies the property red as that property, without conveying information about the colour of the apple beyond the fact that it is the property sensuously detected or represented. The author notes that this kind of pristine (i.e. extremely basic) attribution is not to be found in overt languages (English, ...) because it is too basic: it does not serve the purposes of communication, but is merely representational – this restriction, however, is perhaps too extreme. As regards the issue of what Mentalese sentences are actually made of, the author is more inconclusive: he can only assume that “for a Mentalese term to be meaningful, it must be made of some as yet unspecified physical matter”. The book finishes with an analysis of the relationship between Sententialism and consciousness (chapter 7).
While not truly innovative, this book provides a very clear approach to Sententialism. It is both technical and highly accessible: the author regularly takes examples to make his point and poses the questions to build on before attempting to answer them. Moreover, the table of contents contains a very short summary of the main issues for each chapter. The book should be read more as a thought-provoking frame, maybe, than as definite findings propped up by evidence. A lot of the conclusions and assumptions are only hypotheses, introduced by many phrases such as “probably”, “maybe”, “let us suppose that...” or “it is not utterly wild to postulate that...”. The author is aware of this shortcoming, which he attributes to his object of inquiry: he warns that “philosophical books often fail to establish conclusively what they promise, and what they do manage to demonstrate is typically tenuous”. Still, Sententialism is a consistent construction and raises questions about the relationship between thinking and overt language: for instance, if Mentalese there is, does it follow the fundamentals of Universal Grammar?
The theory also needs to be tested against more complex mental operations than pristine attribution: some of our more elaborate thinking seems not to be performed through words. The book will be of particular interest as well to cognitivists who deal with categorisation or with the notion of “mental representation” created by overt discourse: although these notions are not mentioned by the author, the discussion of Mentalese and networks of beliefs raises the question of the nature of mental representations and of what exactly discourse makes accessible in context.
Cercles © 2011