Paul Needham

From a modern point of view, compounds are contrasted with elements of which they are composed, and the two categories combine to give the category of substances. Mixtures, on the other hand, might be understood to contrast with pure substances (substances in isolation), so that mixtures are quantities of matter containing several substances (be they compounds or uncombined elements) whereas pure substances are understood to be quantities of matter exhausting the material contents of a region of space which contain only one substance, and might be either a compound or an element. But heterogeneous quantities of matter (comprising several phases) might also be understood to be mixtures. Thus, it is common to speak of a mixture of ice, liquid water and water vapour in equilibrium at the triple point of water, although there is only a single substance present. The general concept of mixture as treated by Gibbs’ phase rule is a quantity of matter comprising either several substances or several phases, “or” being used in the inclusive sense. Gibbs’ phase rule is a relatively recent development in the history of the concept of chemical substance, dating from the latter half of the nineteenth century. The concepts and distinctions it presupposes have not always been apparent. I trace some of the important conceptual developments underlying the important distinctions at issue here from Aristotle, who didn’t distinguish either substance and phase, or what are now called compounds and solutions.

The Aristotelian notion of a proper mixture is that of a homogeneous body potentially separable into a definite proportion of elements. Its relation to more modern chemical ideas is not without interest despite the success of modern atomic theory. But there is a fundamental conflict entailed by Aristotle’s two approaches to the characterisation of elements, one in terms of the properties they exhibit in isolation and another in terms of their role as constituents of compounds. Although one source of problems for Aristotle has been put aside in modern chemistry, it seems that the underlying tension remains an issue.

The accumulation of knowledge concerning the character and transformations of substances from ancient times constitutes progress in chemistry, which has accelerated enormously since the end of the 17th century. The present short article focuses on some themes in the development of theorising and conceptual clarification at the macroscopic and microscopic levels during the 19th and 20th centuries. This covers the general understanding of substances in relation to phase and the general notion of a mixture, the distinction between daltonides and berthollides, and issues connected with the notion of molecular structure. Together with the enormous body of factual material accumulated particularly over the last two centuries, this is progress in chemistry. The epistemological approach seems to give the best account of it amongst philosophical views on progress currently on offer.

Chemistry builds on distinctions of substance, which presupposes that matter can be divided into substances and compared with other matter and itself on different occasions as being of the same substance. Even identifying a quantity of matter as comprising a single substance presupposes the same substance relation, it being a quantity all of whose spatial parts are the same substance. But criteria of purity have been important for isolating substances and investigating their characteristic properties, which can in turn be used for establishing sameness of substance. With the development of chemistry into a theoretical science it became important that such criteria and characteristics should have a systematic theoretical basis. Thermodynamics was, perhaps, the first comprehensive theory to systematically divide the mass of the bodies with which it deals into distinct substances and offer general criteria governing the number of substances present. But the applicability of such macroscopic criteria is restricted to equilibrium conditions on a macroscopic time scale. They can be compared with microscopic conceptions of molecular structure, with which they have been complemented—some would say superseded. Since many substances are not molecular, however, the general formulation of a microscopic sameness of substance criterion remains unclear.

Prior left us with a problem which he stated in the following way: ‘Very roughly, it would seem that countable “things” are made or grow from bits of stuff, or from other countable “things”, that are already there. The precise logic of this process hasn’t been worked out yet, and until it has been, it seems likely that any tensed predicate logic can only be provisional in character.’ Although I disagree with much of the philosophy of time underlying Priorean tense logic, the problem of presenting a coherent view of ‘countable things’ and ‘bits of stuff’ within a framework accommodating temporal reference and modal expressions along broadly the lines Prior proposed remains an interesting one. The present paper sketches a general framework within which more specific questions raised by this general issue can be addressed, and makes a number of tentative suggestions without, however, proceeding so far as to settle on and develop a definite logical theory.