sensations are assumed to be? And, after
all, was not Kant a bit premature in proclaiming the _finality_ of his
analysis and of his refutation of empiricism for all time? The searching
question, Why should the future resemble the past? had received no
answer, and so might not the mind itself, with all its categories, be
susceptible to change? Was it certain that the miracle whereby the data
presented to our faculties conformed to them would be a standing one?
Had not Kant himself as good as admitted that our faculties might
distort reality instead of making it intelligible?

The truth is that at this point Kant is open to a charge against which
the assumptions he shared with Hume admit of no defence. Hume had been
the first to discover that we are in the habit of trying to rationalize
our sense-data by putting ideal constructions upon them, though he had
abstained from sanctifying the practice by a hideous jargon of technical
terminology. But this way of eking out the facts only seemed to him to
_falsify_ them. Truth in his view was to be reached by accepting with
docility the sensations given from without. To set to work to 'imagine'
connections between them, and to claim for them a higher truth, had
seemed to him an outrage. What right, then, had Kant to legitimate the
mind's impudence in tampering with sensations? Was not every _a priori_
form an 'imagination,' and a vain one at that?

To these objections the Kantian school have never found an answer. They
have simply repeated Kant's phrases about the necessary
'presuppositions' which were to be added to Hume's data. The English
psychologists (the Mills, Bain, etc.) exhibited a similar fidelity. They
never accepted the _a priori_, but relied on 'the association of ideas'
to build up a mind out of isolated sensations. But was this expedient
really thinkable? For if all 'sensations' or qualities are separate
entities, how can the addition of more 'distinct existences' of the same
sort really bind them together? If in 'the cat is upon the wall,' 'upon'
is a distinct entity which has to relate 'cat' and 'wall,' what is to
connect 'cat' with 'upon' and 'upon' with 'wall'? The atomizing method
carried to its logical extreme demands that not only 'sensations' but
also 'thoughts' should be essentially disconnected, and then, of course,
_no_ thinking can cohere.

Psychology, then, had worked itself to a breakdown by accepting the
'sensationalistic' analysis offered by Hume