Source
Botvinick & Cohen (1998), LIT-1277, Nature 391:756, a one-page scientific correspondence, read in full in NOTE-1006: the text, both figures and their captions. The copy read is the publisher’s typeset page on a third-party host, as the NOTE’s history says.
What was actually shown
The illusion. Each subject’s left hand was hidden behind a screen and a life-sized rubber left hand was placed in view. Both were stroked with paintbrushes, “synchronising the timing of the brushing as closely as possible”. After ten minutes, ten subjects rated nine statements in random order. The three predicted ones were affirmed: that touch was felt where the rubber hand was touched, that it was caused by the brush on the rubber hand, and “I felt as if the rubber hand were my hand” (P<0.002, Fig. 1). Five of the six control statements had means on the disagree side. Eight of ten subjects used terms of ownership unprompted.
The drift. After thirty minutes of exposure, blind intermanual reaches toward the real hand were displaced toward the rubber hand. The displacement grew with the reported share of the exposure during which the illusion was felt (r² 0.47, p<0.03, about ten points).
The control. With “a small asynchrony” the illusion was reported for 7% of the exposure against 42%, and reaches moved 13 mm away from the rubber hand against 23 mm toward it in the synchronous group. This could have come out otherwise: seeing a hand brushed beside one’s own, at any timing, could have produced the same reports and drift. It did not.
What this does not say
- It does not separate synchrony from expectation. The predicted items were known to the experimenters, the intended answer is guessable from the set-up, and the paper does not discuss demand characteristics. The control items guard against blanket agreement, not against suggestion specific to the touch and ownership items. That is what the promotion test asks for.
- It does not give effect sizes or the control group’s size. The tests behind the P-values are not named. Figure 1’s corrected threshold (P<0.018) is looser than the uncorrected one (P<0.002), which the paper does not explain.
- It does not say what the drift measures. The drift may be ownership or only visuo-proprioceptive recalibration. It was measured after a longer exposure than the questionnaire, and the paper ties the two only through the reported duration.
- It does not establish the mechanism. The authors’ constraint- satisfaction account rests on a connectionist model that is neither shown nor cited. The predictive accounts filed alongside this one, THEORY-769 and THEORY-767, are explanations of this finding, not part of it.
- It is about a hand. The conclusion that intermodal matching “can be sufficient for self-attribution” concerns a body part. Whether ownership of the whole body, self-location or a self works the same way is THEORY-768’s question, and Blanke & Metzinger (LIT-1274) argue that it does not.
Connections
No relation is declared: no open question asks how body ownership arises.
- NOTE-372, claim C6. Metzinger’s “body ownership can be manipulated by multisensory conflict” rested on studies the record did not hold. This is the primary result that work extends.
- THEORY-458 and LIT-1009. The rubber hand gives ownership of a hand that does not move, by passive stimulation alone. That is consistent with the split between ownership and agency, but it does not test THEORY-458’s comparator half.
- THEORY-064. Its promote_when already says that more body-ownership illusions show the self-model can be manipulated, not that it is all a self is. This finding is of that kind.
- QUESTION-048. A self-report measure is paired with a behavioural measure that covaries with it, though the two were not taken on the same trials.