Source
Ten works, each read in full:
- The founding report: Botvinick & Cohen (1998), LIT-1277, NOTE-1006.
- Replications with named tests:
- Drift: Rohde, Di Luca & Ernst (2011), LIT-1281, NOTE-1009, and Abdulkarim & Ehrsson again.
- Suggestion:
What was actually shown
The report difference replicates. Botvinick & Cohen’s ten subjects affirmed referred touch and ownership after synchronous brushing. Longo et al. found ownership and location crossing from negative to positive with synchrony, within subjects. Kalckert & Ehrsson found the same with a moved, not brushed, model hand: a delay of about 500 ms removed it. Abdulkarim & Ehrsson found ownership depended on synchrony in two experiments of 18 (d 0.58–1.76). In Lush et al.’s 353 participants, as reanalysed by Ehrsson et al., the synchronous-minus-asynchronous difference in illusion ratings was significant in every hypnotisability quartile and unrelated to hypnotisability (rho = −0.066, BF01 = 5.7).
The ownership report is the weak one. In Lush et al.’s sample the ownership item averaged 0.7 on a −3 to +3 scale, against 1.9 and 1.2 for the two referred-touch items. Agreement with it rose 0.76 per point of hypnotisability and reached minimal agreement only in the top 31% (Lush & Seth). In Roseboom & Lush’s resampling of the synchronous reports into studies of 20, samples from below the median of hypnotisability agreed with the ownership item about 4% of the time. They agreed with the referred-touch item 95% of the time. The ownership result held at n = 50.
Drift comes apart from the reports. In Rohde et al., drift toward the rubber hand was as large when the hand was only seen, with no brushing, as under synchronous brushing (about 6 cm). It was as large under asynchronous brushing interrupted every 10 s. Only continuous asynchrony for 120 s kept it near zero, while ownership by questionnaire needed synchrony. In Abdulkarim & Ehrsson, a motor moved the hidden hand 8 cm toward or away from the rubber hand. Pointing followed it by about 5–6 cm, and ownership did not change with the direction of movement. This meets the second refutation condition of THEORY-765, which this THEORY corrects.
The standard controls do not equalise expectations. People shown the procedure but never stimulated expected the illusion statements more than the control statements (d = 1.40). They also expected them more under synchronous than asynchronous brushing (d = 1.05, n = 20; Lush 2020). In Lush et al.’s sample, expectations were higher for synchrony, and that gap did not vary with hypnotisability (Lush & Seth). An instruction that asynchrony works best did not change expectations. Both sides call that manipulation a failure.
What this does not say
- It does not say the stimulation causes the report difference. Ehrsson et al. read the difference’s independence from hypnotisability as evidence that it is perceptual. Lush & Seth show that uniform unequal expectations predict the same null. Neither side has run the experiment that would decide it, with expectations matched in stimulated participants. That is the promotion test.
- It does not say the reports are suggestion. Lush (2020) says plainly that his result is not evidence that the illusion is suggestion. Lush et al.’s pre-registered test of hypnotisability against the difference score was inconclusive.
- It does not say drift is unrelated to ownership. With the hand still, the classic drift and its correlation with ownership reports came back (Abdulkarim & Ehrsson, r = .562). Longo et al. found ownership and location each predicted drift. The claim is that drift is not a measure of the reports: it can be produced without them and moved without moving them. Both drift studies are small (groups of 10 and 18), and Abdulkarim & Ehrsson’s null has no equivalence test.
- It does not settle what the ownership item measures. Longo et al.’s components and Kalckert & Ehrsson’s dissociations rest on the same questionnaire logic the suggestion papers question.
- It is about a hand. Whole-body ownership is THEORY-768’s question. The suggestion papers predict, without testing, the same confound for the full-body illusion.
Connections
No relation is declared to a question: no open question asks how body ownership arises.
- THEORY-765 is corrected. Its drift clause failed its own refutation condition, and its suggestion control was too weak to tell the two readings apart.
- THEORY-769 and THEORY-767 explain this report difference as inference. Both now need the expectation control as well. Their source authors, Tsakiris and Seth, sign opposite sides of the 2022 exchange.
- THEORY-458. Kalckert & Ehrsson’s passive condition gives ownership reports without agency reports, and their rotated hand gives agency without ownership. Both rest on reports of the kind this THEORY treats as unsettled.
- NOTE-372, claim C6. Ownership reports can be manipulated, but the record no longer holds that multisensory conflict, rather than expectation, is what manipulates them.
- QUESTION-048. A worked case: report and behaviour dissociate, and both sides agree on the numbers while disagreeing on what they show.
Narrowed and promoted (2026-10-20)
The promotion audit (batch 075) promoted this entry under ADR-047 v3. Its title already rested on ten works read in full and claimed no causal attribution; the clause on expectations is added so that “raises reports” cannot be read as saying the stimulation produces the difference.
That reading, Ehrsson et al.’s claim that the illusion is perceptual and multisensory (NOTE-997, C8), went to THEORY-827 under ADR-047 v4, with this entry’s former promote_when, which is its test.