Handedness doesn't appear at the same time as language - though Corballis (.
http://www.bbsonline.org/Preprints/Corballis/Referees/) believes that right handedness may have evolved due to 'left-hemispheric dominance for vocalization' (not quite the same thing). I believe in Carlson et al (2004) it maintains that hand dominance can be observed in the womb - don't have the book handy right now, but I think it was bloody early too. However, found something interesting about animals and handedness...
[/quote]Reply to Hopkins and Cantalupo: Chim-panzee Right-Handedness Reconsidered—Sampling Issues and Data PresentationA. Richard Palmer*Department of Biological Sciences, University of Alberta,Edmonton, Alberta T6G 2E9, CanadaGENERAL ISSUES REGARDING PRIMATELATERALITY STUDIESHopkins et al. (2001) deserve much credit for hav-ing raised the bar for laterality studies in nonhumanprimates by routinely studying such a large numberof individuals and by developing well-defined tasks,such as the TUBE task, to try to quantify individuallaterality in a reliable way. They also deserve creditfor repeating prior studies to assess the consistencyof results, because true replications are so rarelyconducted (Palmer, 2000). Nonetheless, the reply byHopkins and Cantalupo raises several importantissues regarding their and others’ work on chimpan-zee handedness.First, Hopkins and Cantalupo are correct that theunexpected patterns revealed by funnel graphs ofthe 1994 handedness data (Hopkins, 1994; Palmer,2002) only became apparent because sample size(number of handedness observations per individual)varied. They are also correct that standardizing thenumber of observations per individual totally elim-inates any possible sample-size effects: a funnelgraph is only useful after the fact, as a form ofquality-check of the data. The real question, how-ever, is not as Hopkins and Cantalupo suggest,“How many observations to take per individualchimpanzee?,” but rather, “What aspects of the be-havioral sampling protocol gave rise to the unex-pected patterns in the first place?”Eliminating variation in the number of observa-tions per individual merely eliminates the symp-toms, but not the underlying cause(s), of the unusualdistributions reported in Palmer (2002): why wasright-handedness of individual chimpanzees morepronounced among those individuals for whichfewer observations were recorded? If the underlyingcauses were not eliminated in the present study asindicated by Hopkins and Cantalupo, then they maystill contribute to an apparent population-levelright-handedness. Some further reflections by Hop-kins et al. on the possible underlying causes of theodd patterns in the 1994 data might provide valu-able insights that would help improve the design offuture primate handedness studies.Second, my critique (Palmer, 2002) was also in-tended to illustrate how graphical presentation ofdata in as unreduced a form as possible (e.g., scat-terplots, frequency distributions) is preferable tostatistical summaries. Tabulated statistical summa-ries, although economical in terms of journal space,invariably obscure details about the data that mayaffect confidence in the results or that might suggestalternative interpretations. If, as was done in theoriginal 1994 study, Hopkins and Cantalupo were topublish the raw data behind the statistical summa-ries tabulated in their reply and also in the moreextensive study of Hopkins et al. (2001), this wouldallow the evidence for population-level right-hand-edness (at least among chimpanzees at the YerkesPrimate Research Center (YRPRC)) to be judgedmore fully and fairly.For example, Hopkins kindly provided to me rawdata on the TUBE task for 109 chimpanzees at theYRPRC (from Hopkins et al., 2001), where the over-all percent right-hand use was very similar to thatreported in Hopkins (1994) for bimanual feeding(Table 1). Seventy-five of these individuals were alsoincluded in the 1994 study, and of these, handednesswas based on more than 25 observations for 56 ofthem (for the justification to exclude individualswith 25 or fewer observations, see Palmer, 2002).The correlation between percent right-hand use dur-ing bimanual feeding (holding food in one handwhile removing portions with the other) and percentright-hand use by the same individual in the TUBEtask (holding a tube in one hand and scooping outpeanut butter with a finger of the other) is not overlycompelling (Fig. 1). Although these tasks are some-what different, I am surprised to see so little consis-tency between tasks that require the simultaneoususe of both hands in such a similar, coordinatedfashion during feeding. If chimpanzees at the YRPRCexhibit population-level right-handedness for bi-manual tasks, as the data in Table 1 suggest, ratherlittle of this right-handedness appears to derivefrom consistent hand preferences by individualchimpanzees.Finally, as someone outside the field of primatelaterality, I have always been troubled by handed-ness data obtained from captive animals, as haveother primatologists (McGrew and Marchant, 1997).No matter how much care is taken to avoid intro-ducing handedness biases in the sampling protocolGrant sponsor: Natural Sciences and Engineering Research Councilof Canada; Grant number: A7245.*Correspondence to: A. Richard Palmer, Department of BiologicalSciences, University of Alberta, Edmonton, Alberta T6G 2E9, Canada.E-mail:
[email protected]. 1 July 2002; accepted 6 August 2002.DOI 10.1002/ajpa.10177AMERICAN JOURNAL OF PHYSICAL ANTHROPOLOGY 121:382–384 (2003)©2003 WILEY-LISS, INC.
--------------------------------------------------------------------------------
Page 2
of a particular study, I wonder to what degree cap-tive chimpanzees have been irreconcilably biasedtoward right-handedness simply by watching theirstrongly right-handed captors go about many otherdaily tasks unrelated to the study. Chimps may notonly imitate human behaviors, or behaviors of otherchimps; they may also imitate the overwhelminghuman tendency to use the right hand for mosttasks. I’m afraid I see little solution to this problembut to encourage more studies of wild populations,like those of Marchant and McGrew (1996) andMcGrew and Marchant (2001).TECHNICAL ISSUES REGARDING HOPKINSAND CANTALUPOTo avoid confusion, I should note that Hopkins andCantalupo misrepresent one expected pattern re-vealed by a funnel graph. No “basic statistical assump-tions of increasing effects with increasing sample size”are made when interpreting a funnel graph. As out-lined in my original paper (Fig. 1 of Palmer, 2002), thevariability should actually decrease with increasingsample size, and the expected mean effect size shouldbe independent of sample size.Hopkins and Cantalupo are correct that the numberof observations per individual chimpanzee in the 1994study was confounded by age (i.e., more observationswere obtained from younger individuals, and youngerindividuals tend to be less lateralized), and that this atleast partly contributed to the decline in right-hand-edness with increasing number of observations perindividual, as I too noted (Palmer, 2002). However,this pattern was even more pronounced among olderchimpanzees (see Table 3 of Palmer, 2002), so thepatterns revealed by the funnel graphs were not duesolely to the confounding effects of age. The questiontherefore still remains: why was right-handednessmore pronounced among individuals for which fewerobservations were recorded?Finally, as I understand their methods, the re-sults reported by Hopkins and Cantalupo appear tobe confounded by pseudoreplication (Hurlburt,1984), and may therefore yield inflated estimates ofindividual hand preference. Equal sample sizeswere obtained for each individual chimpanzee byrecording “the first 20 hand-use responses” on fourseparate occasions when performing the TUBE task(Hopkins et al., 2001). However, if an individualchimp holds the tube in its left hand, and inserts itsright finger into the tube 20 times in succession, is itappropriate to score this as 20 independent inser-tions of the right finger (as Hopkins and Cantalupoappear to have done), or should it simply be scoredas one grasp of the tube with the left hand? If eachfinger insertion was scored as an independent obser-vation, I am not surprised that “the majority (90%)of chimpanzees show a significant hand preference”according to this measure (note, however, that theactual data reported in Hopkins and Cantalupo(2002) show only 100 of 132 chimpanzees (or 76%)TABLE 1. Frequencies of hand use by chimpanzees at YRPRC reported in different studiesNumber of individuals% Right2Activity (source)LeftAmbilateral1Right28585465.9%Bimanual feeding (Hopkins, 1994)32195964.8%TUBE task (Hopkins, 1995)33225462.1%TUBE task (Hopkins et al., 2001)29327171.0%TUBE task (Hopkins and Cantalupo, 2003)1Ambilateral means difference in hand use between sides in an individual did not exceed that expected due to binomial samplingvariation.2Percent of those individuals exhibiting significant handedness.Figure 1. Consistency of handedness (raw percent right-hand use) for two related tasks among 56 individual chimpanzeesat YRPRC: bimanual feeding (holding a food object in one handwhile removing portions of it with the other, from Hopkins, 1994)and TUBE task (holding a tube in one hand and scooping outpeanut butter with a finger of the other, from Hopkins et al.,2001). Line represents least-squares linear regression fit to thedata. The association, although positive, is not significant statis-tically, either for all individuals (r0.12, P0.36, Spearmancoefficient of rank correlation) or when restricted to individualsthat were already adolescents and adults in the 1994 study (N47, r0.21, P0.16). For reasons outlined in Palmer (2002),only individuals with more than 25 hand-use observations in the1994 data were included. Handedness for all individuals in theTUBE task was based on more than 25 observations, so nonewere excluded. Age groupings among chimps in the 1994 studywere: juvenile ( 7 years old), adolescent (7–15 years old), adult( 15 years old).CHIMPANZEE RIGHT-HANDEDNESS RECONSIDERED383
--------------------------------------------------------------------------------
Page 3
exhibited significant handedness). Might this alsoaccount for the somewhat higher percent right-hand-edness reported in Hopkins and Cantalupo comparedto Hopkins et al. (2001) (Table 1)? Perhaps Hopkinsand Cantalupo could report the number of times achimp picked the tube up, or rotated the tube to accessthe other end with the same hand, or switched handsused to extract the peanut butter with the other hand.These behaviors would seem to provide better inde-pendent measures of hand preference.ACKNOWLEDGMENTSI thank Bill Hopkins for providing me the rawdata from his 2001 study, and L. Hammond forhelpful comments on the manuscript. My researchprogram has been supported by sustained fundingfrom the Natural Sciences and Engineering Re-search Council of Canada (operating grant A7245).LITERATURE CITEDHopkins WD. 1994. Hand preferences for bimanual feeding in 140captive chimpanzees (Pan troglodytes): rearing and ontogeneticdeterminants. Dev Psychobiol 27:395–407.Hopkins WD. 1995. Hand preferences for a coordinated bimanualtask in 110 chimpanzees (Pan troglodytes): cross-sectional anal-ysis. J Comp Psychol 109:291–297.Hopkins WD, Fernandez-Carriba S, Wesley MJ, Hostetter A,Pilcher D, Poss S. 2001. The use of bouts and frequencies in theevaluation of hand preferences for a coordinated bimanual taskin chimpanzees (Pan troglodytes): an empirical study compar-ing two different indices of laterality. J Comp Psychol 115:294–299.Hurlburt SH. 1984. Pseudoreplication and the design of ecologicalfield experiments. Ecol Monogr 54:187–211.Marchant LF, McGrew WC. 1996. Laterality of function in wildchimpanzees of Gombe National Park: comprehensive study ofspontaneous activities. J Hum Evol 30:427–443.McGrew WC, Marchant LF. 1997. On the other hand: currentissues in and meta-analysis of the behavioral laterality of handfunction in nonhuman primates. Yrbk Phys Anthropol 40:201–232.McGrew WC, Marchant LF. 2001. Ethological study of manuallaterality in the chimpanzees of the Mahale mountains, Tan-zania. Behaviour 138:329–358.Palmer AR. 2000. Quasireplication and the contract of error:lessons from sex ratios, heritabilities and fluctuating asymme-try. Annu Rev Ecol Syst 31:441–480.Palmer AR. 2002. Chimpanzee right-handedness reconsidered:evaluating the evidence with funnel plots. Am J Phys An-thropol 118:191–199.384A.R. PALMER
Hope the mods can clean this UP