In his second article he abandons this interpretation, being convinced that Marx's ideal construction does not aim at supplying a complete explanation of the phenomena of economics by means of the increasing complexity of his combinations. And, in my opinion, he did well to abandon it; not only for the excellent reason stated by him, that Marx's inquiry does not include an entire system of economics, but also because the process suggested by him does not explain why Marx, in analysing the economic phenomena of the second or third sphere, ever _used concepts whose place was only in the first one_. It does not explain what I have called the _eliptical comparison_, and herein lies the difficulty of Marx's work, or rather of the literary statement of his thought. If the correspondence between labour and value is only realized in the simplified society of the first sphere, why insist on translating the phenomena of the second _into terms of the first_? Why give the name transformation of surplus value to what makes its appearance as the natural economic result of capital which must have (from its very nature as capital) a profit? Does Marx offer an explanation connecting ground and consequence, or does he not rather draw a _parallel between two different phenomena_, by which the diversities illuminating the origins of society are set in relief? But Sorel now advances to precisely this conclusion, borrowing a happy phrase from his first article: that Marx's work is not intended to explain by means of laws analogous to physical laws, but only to throw partial and indirect light on economic reality. The method which Marx employs in his inquiry, says Sorel, is a _metaphysical instrument_; he makes a _metaphysics of economics_. This expression may be satisfactory or not, according to the different meanings given to the word _metaphysics_; but the idea is accurate and true. Marx builds an ideal construction which helps him to explain the conditions of labour in capitalist society.