The article attempts to settle a controversy between d f pears and j o urmson over the nature of russell's early theory of memory. it is shown that contrary to what pears claims in his "bertrand russell and the british tradition in philosophy," russell had explicitly abandoned a realist account of memory by 1915. the article sides with urmson as against pears, but apparently both have overlooked two of russell's little noticed 1915 papers in the "monist.".

Urmson is correct in holding that russell's use of "logical fiction" does usually involve ontological implications. But the issue is more complex than he seems to realize. Because russell's program of logical construction is revisionary, The question "are there really x's?" is ambiguous and can be taken as asking either: (a) are there x's as thought of pre-Analytically? or (b) are there x's as thought of post-Analytically? russell gives different answers in each case.

J. o. urmson's contention that russell held that 'to show that 'x' is an incomplete symbol is tantamount to showing that there are no x's' is shown to rest partly upon a misreading of "principia", pp. 71-72, where russell reveals what he means by a 'definite proof' that a symbol is incomplete, and partly upon a misunderstanding of russell's use of the expression 'logical fiction'.