“But surely in case 14c) B acted entirely automatically. If all that happened was really what was described there, he did not know why he chose the bolt he did choose. He had no reason for choosing it. If he chose the right one, he did it as a machine might have done it”. Our first answer is that we did not deny that B in case 14c) had what we should call a personal experience, for we did not say that he didn't see the materials from which he chose or that which he chose, nor that he didn't have muscular and tactile sensations and such like while he did it. Now what would such a reason which justified his choice and made it non-automatic be like? (i.e.: What do we
17.
imagine it to be like?) I suppose we should say that the opposite of automatic comparing, as it were, the ideal case of conscious comparing, was that of having a clear memory image before our mind's eye or of seeing a real sample & of having a specific feeling of not being able to distinguish in a particular way between these samples and the material chosen. I suppose that this peculiar sensation is the reason, the justification, for the choice. This specific feeling, one might say, connects the two experiences of seeing the sample, on the one hand, and the material on the other. But if so, what connects this specific experience with either? We don't deny that such an experience might intervene. But looking at it as we did just now, the distinction between automatic and non-automatic appears no longer clear-cut and final as it did at first. We don't mean that this distinction loses its practical value in particular cases, e.g., if asked under particular circumstances, “Did you take this bolt from the shelf automatically, or did you think about it?”, we may be justified in saying that we did not act automatically and give as a reason || explanation we had looked at the material carefully, had tried to recall the memory image of the pattern, & had uttered to ourselves doubts and decisions. This may in the particular case be taken to distinguish automatic from non-automatic. In another case however we may distinguish between an automatic & a non-automatic way of the appearance of a memory image, and so on.