My dear M^r. Darwin,

The MS which I return with best thanks is in the highest degree interesting, & I cannot refrain from saying how glad I am that, like a worm, you have got hold of yet another so good a leaf to drag into your wonderful store.¹

Concerning the question of intelligence, it seems to me that the difficulty arises from the fact of its being a thing, which like all other things, is gradually evolved. The line, therefore, between non-intelligent and intelligent adjustment is, as we should à priori expect, as difficult to draw as it is, say between an egg and a bird. This consideration, however, does not of course meet the requirement that we feel for some criterion of intelligence when sufficiently evolved to be recognized as such. What we want is a test that may be taken as a line—even though an artificial one—to divide actions which we agree to call intelligent from those which we agree to call non-intelligent. Such a test, it seems to me, can only be furnished by the question—Does the animal learn by its own individual experience? For the test must be, as you say, an objective one; we cannot get inside an animal’s mind so as to obtain direct, or subjective, knowledge of its operations. And, if the test is to be objective, I see no analytical filter that can strain off all the class of non-intelligent adjustments due to heredity, save the question above stated.

Observe, I do not doubt that before this test is conformed to in the ascending series of psychological life, intelligence has begun to dawn & may be tolerably far advanced; so that if we could obtain any subjective test we should probably be able to draw our line of demarcation lower down in the series. But as we are shut up to an objective test, it appears to me that the best thing we can do if we aim at certain definitio⁠⟨⁠n⁠⟩⁠ is to make our distinction—arbitrary though it may be—at the place where we are first sure that there is a distinct thing to define.

“Thems my sentiments.”² But I know they will not do much to meet the case of earthworms; for the actions of these animals seem to stand just on the border-land. If, however, it could be shown by experiment—which I think not at all impossible—that a particular earthworm admits of being taught by experience how best to manipulate some awkward exotic leaf, so that after finding out the best way of manipulating it would afterwards always or generally choose this way— if such could be shown to be the case, there could no longer be any question as to the action being in the full sense of the word intelligent. On the other hand, if experiment failed to show this, all, I think, that could be said would be that the action, if intelligent, is not intelligent in a degree sufficiently high to admit of our certainly classifying it as such.³

Turning now to what you give in the MS as the “chief element in an act that deserves to be called intelligent”, I think it would be well to avoid hyper-criticism of the Mivart class, to guard the statement so as to make it evident that you do not mean self-consciousness.⁴ An animal, or a man, may certainly perform an intelligent act without at all thinking about his own thoughts or of anything that is passing in his own mind. Of course I know that you mean consciousness & not self-consciousness. But this, after all, is only re-stating the difficulty. There clearly can be no intelligence without consciousness,—just as clearly as there may be intelligence without self-consciousness. But forasmuch as consciousness—or the power however incipient of distinguishing between pleasure & pain—must be supposed to dawn before the advent of intelligence of which it is the necessary condition, it follows that we gain no help in defining intelligence as that which presents the element of consciousness.

Probably most reflex actions & non-intelligent hereditary habits have been elaborated only in virtue of a consciousness—however rudimentary—of pleasure & pain.

I do not suppose that this dis-quisition will be of any use to you; but when I begin to write I am apt to run on regardless of the patience of anyone who manages to get to the end. Therefore this letter has grown much longer than I at first intended, which explains the half-sheets on which it is written.

Again thanking you very much for letting me see the MS, & also for the notes of reference⁵ | I remain, | Very sincerely & most respectfully yours. | Geo. J. Romanes.