Sismondi thus grounds his views in a theory of income. What is income, and what is capital? He pays the greatest attention to this distinction which he calls 'the most abstract and difficult question of political economics'. The fourth chapter of his second book is devoted to this problem. As usual, Sismondi starts his investigation with Robinson Crusoe. For such a one, the distinction between capital and income was still 'confused'; it becomes 'essential' only in society. Yet in society, too, this distinction is very difficult, largely on account of the already familiar myth of bourgeois economics, according to which 'the capital of one becomes the income of another', and _vice versa_. Adam Smith was responsible for this confusion which was then elevated to an axiom by Say in justification of mental inertia and superficiality. It was loyally accepted by Sismondi. 'The nature of capital and of income are always confused by the mind; we see that what is income for one becomes capital for another, and the same object, in passing from hand to hand, successively acquires different denominations; the value which becomes detached from an object that has been consumed, appears as a metaphysical quantity which one expends and the other exchanges, which for one perishes together with the object itself and which for the other renews itself and lasts for the time of circulation.'[159] After this promising introduction, Sismondi dives right into the difficult problem and declares: all wealth is a product of labour; income is part of wealth, and must therefore have the same origin. However, it is 'customary' to recognise three kinds of income, called rent, profit and wage respectively, which spring from the three sources of 'land, accumulated capital and labour'. As to the first thesis, he is obviously on the wrong tack. As the wealth of a society, i.e. as the aggregate of useful objects, of use-values, wealth is not merely a product of labour but also of nature who both supplies raw materials and provides the means to support human labour. Income, on the other hand, is a concept of value. It indicates the amount to which an individual or individuals can dispose over part of the wealth of society or of the aggregate social product. In view of Sismondi's insistence that social income is part of social wealth, we might assume him to understand by social income the actual annual fund for consumption. The remaining part of wealth that has not been consumed, then, is the capital of society. Thus we obtain at least a vague outline of the required distinction between capital and income on a social basis. At the very next moment, however, Sismondi accepts the 'customary' distinction between three kinds of income, only one of which derives exclusively from 'accumulated capital' while in the other two 'land' or 'labour' are conjoined with capital. The concept of capital thus at once becomes hazy again.