In 1904 appeared the work on this subject which has attracted the most attention—R. Semon's Die Mneme.[514] This was an elaborate treatment of the question from the materialistic point of view, the main assumption of Semon's theory being that the action of a stimulus upon the organism leaves a more or less permanent material trace or "engramm," of such a nature as to modify the subsequent action of the organism.
Applied to the explanation of heredity and development, Semon's theory comes to very much the same as Weismann's, with engramms substituted for determinants, but it has the great advantage of allowing for the transmission of acquired characters. The application of the concept of stimulus is valuable and suggestive, but it seems to us that the memory theory of heredity can be properly utilised only by adopting a frankly Lamarckian and vitalistic standpoint, and this standpoint Semon expressly combats. As Ward[515] points out in his illuminating lecture on heredity and memory—"Records or memoranda alone are not memory, for they presuppose it. They may consist of physical traces, but memory, even when called 'unconscious,' suggests mind; for, as we have seen, the automatic character implied by this term 'unconscious' presupposes foregone experience.... The mnemic theory then, if it is to be worth anything, seems to me clearly to require not merely physical records or 'engrams,' but living experience or tradition. The mnemic theory will work for those who can accept a monadistic or pampsychist interpretation of the beings that make up the world, who believe with Spinoza and Leibniz that 'all individual things are animated albeit in divers degree'" (pp. 55-6).
Perhaps the best and most ingenious treatment of memory and heredity from a physical standpoint is that offered by E. Rignano in his book, Sur la transmissibilité des caractères acquis.[516] Rignano seeks to construct a physico-chemical "model" which will explain both heredity and memory.
His system, which is based more firmly upon the facts of experimental embryology than Semon's, postulates the existence of "specific nervous accumulators." The essential hypothesis set up is that every functional stimulus is transformed into specific vital energy, and deposits in the nucleus of the cell a specific substance which is capable of discharging, in an inverse direction, the nervous current which has formed it, as soon as the dynamical equilibrium of the organism is restored to the state in which it was when the original stimulus acted upon it. These specific nuclear substances, different for each cell, are accumulated also in the nuclei of the germinal substance, constituting what Rignano calls the central zone of development. That is to say, each functional adaptation changes slightly the dynamical equilibrium of the organism, and this change in the system of distribution of the nervous currents leads to the deposit in the central zone of development of a new specific substance. In the development of the next individual this new specific element enters into activity, and reproduces the nervous current which has formed it, as soon as the organism reaches the same conditions of dynamical equilibrium as those obtaining when the stimulus acted on the parent.
Development can thus be regarded as consisting of a number of stages, at each of which new specific elements enter automatically into play and lead the embryo from that stage to the stage succeeding. The germinal substance on this theory of Rignano's is to be regarded as being composed of a large number of specific elements, originally formed as a result of each new functional adaptation, but now forming part of the hereditary equipment.
The theory represents an advance upon the more static conceptions of Semon. It owes much to Roux's influence.
In this country, the mnemic theories have been championed particularly by M. Hartog[517] and Sir Francis Darwin.[518]
[508] The quotations are taken from the 1910 reprint, London, Fifield.
[509] Ueber das Gedächtnis als eine allgemeine Funktion der organisierten Materie, Wien, 1870.
[510] Eng. trans, in E. Hering, Memory, p. 9, Chicago and London, 1913.