Pierre Pica
Laboratoire Structure Formelle du Langage, Université Paris 8. Saint-Denis, FR; Centre national de la recherche scientifique, Unité mixte de recherche. Saint Denis, FR.
Bert Vaux
Department of Linguistics, University of Cambridge. Cambridge, UK.
Jeffrey Watumull
Department of Linguistics, University of Cambridge. Cambridge, UK.
The undersigned, by their signatures, express their support for the general points concerning scientific inquiry and conduct discussed here:
Stephen R. Anderson
Cognitive Science Program, Department of Linguistics, Department of Psychology, Yale University. New Haven, CT, US.
Fred B. Bercovitch
Primate Research Institute and Wildlife Research Center, Kyoto University. Inuyama, Aichi, JP.
Gustaro Beritognolo
Department of Linguistics, UQUÀM. Montréal, QC, CA.
Sylvain Bromberger
Department of Linguistics and Philosophy, Massachusetts Institute of Technology. Cambridge, MA, US.
Noam Chomsky
Department of Linguistics and Philosophy, Massachusetts Institute of Technology. Cambridge, MA, US.
Geoffrey A. Clark
School of Evolution and Social Change, Department of Anthropology, Arizona State University. Tempe, AZ, US.
Florian Engert
Department of Molecular and Cellular Biology, Harvard University. Cambridge, MA, US.
Kleanthes K. Grohmann
Department of English Studies, University of Cyprus. Nicosia, CY.
John Halle
Conservatory of Music, Bard College. Annandale-on-Hudson, NY, US.
Morris Halle
Department of Linguistics and Philosophy, Massachusetts Institute of Technology. Cambridge, MA, US.
Daniel Harbour
Cognitive Science of Language Program, Queen Mary University of London. London, UK.
Stephanie Harves
Department of Linguistics, New York University. New York, NY, US.
Dalina Kallulli
Department of Linguistics, University of Vienna. Vienna, AT.
Peter Kosta
Institute for Slavic Studies, University of Potsdam. Potsdam, DE.
Phyllis C. Lee
Department of Psychology, University of Stirling. Sitrling, UK.
Nirmalangshu Mukherji
Department of Philosophy, University of Delhi. Delhi, IN.
Massimo Piattelli-Palmarini
Cognitive Science Program, Department of Linguistics, Department of Psychology, Department of Management and Policy, University of Arizona. Tucson, AZ, US.
Jean-Yves Pollock
Département d études anglophones, Université Paris Est, Marne la vallée.
Gertjan Postma
Meertens Institute, Netherlands Academy of Sciences. Amsterdam, NL.
Henk van Riemsdijk
Center for Cognition, Tilburg University. Tilburg, NL.
Sidarta Ribeiro
Brain Institute, Federal University of Rio Grande do Norte. Natal-RN, BR.
Ian Roberts
Department of Linguistics, University of Cambridge. Cambridge, UK.
Kenneth Safir
Department of Linguistics, Rutgers University. New Brunswick, NJ, US.
Peter Svenonius
Center for Advanced Study in Theoretical Linguistics, University of Tromsø. Tromsø, NO.
Jess Taubert
Yerkes National Primate Research, Emory University. Atlanta, GA, US.
Alberto Paolo Tonda
Institut des Systèmes Complexes, Centre national de la recherche scientifique. Paris, FR.
Yuri I. Alexandrov
Laboratory of Neural Bases of Mind, Institute of Psychology, Russian Academy of Sciences. Moscow, RU.
_________________________________________________________________
Could any researcher consider the integrity of science, let alone that of the individual researcher, to be respected by a process that allows the raiding (1) of a lab to seize all of its materials (hard drives, records, etc.), the prying (2) into of e-mails, raw data, draft manuscripts, and finances (à la Climategate), and the “unusual” (3) (read as “media-inspired”) intervention of a U.S. Attorney? Could any student not demur at a life in science governed by a process of prosecution and punishment for work either unpublished, even unsubmitted, or corrected prior to publication (4)? The issues raised by these questions, though associated with the investigation of former Harvard Professor Marc Hauser’s research (4), are not particular to his case. The issues are—or should be—of interest to anyone who considers science as free inquiry into human nature and the nature of the universe; they should be deeply disquieted by a process for which “problems involving data acquisition, data analysis, data retention” (4) are actionable offenses.
Science is a dialectic defined by debating the methods of data acquisition, the analyses of those data, and the theory within which the data are acquired and analyzed. Prying into ongoing work reveals, unsurprisingly, disputes, biases, and uncertainties to be resolved prior to any publication or unresolved and unpublished; if published, a study with such unresolved problems would probably not be replicable. So this is a human enterprise: bad decisions can be made; some errors (e.g., in data retention) are inevitable, some even to be welcomed if they evidence theoretical or methodological weaknesses to be strengthened. Indeed, “the clearer perception and livelier impression of truth [is] produced by its collision with error” (5). By this method is science remarkably successful.
Thus to undermine the scientific dialectic in the manner described above, as the investigation of Hauser has, is to undermine the very nature—the very possibility—of experimental science. This—not the facts of Hauser’s alleged misconduct—is the scandal. If only one point is to be made by our writing, this is it. Every scientist, tenured and untenured, young and old, and every student considering entering into science, should greatly fear an environment in which every unpublished fragment of research, every financial record, is open to scrutiny by university officials and government investigators. Our goal in this essay is not to defend or evaluate the published information about Hauser’s case, but rather to use what has been made available to establish the more general point.
In August 2010, Harvard University disclosed that it had found Hauser responsible for eight instances of scientific misconduct (4). Their statement neither provided a definition of scientific misconduct, nor discussed the nature of Hauser’s alleged violations, beyond stating that three involved published studies, and the remaining five involved work either unpublished (some unsubmitted) or corrected prior to publication. (We return to these below.) The only relevant definition of scientific misconduct—which Harvard does not follow in its details—is stated in the regulations of the U.S. Office of Research Integrity (ORI): “falsification, fabrication, or plagiarism” (6). This definition is not universally accepted or even universally understood (7). (The seemingly arbitrary definitions of and distinctions between “research misconduct” and “questionable research practice” and normal science are anything but clear, and therefore dangerous.) And due to an ongoing investigation by the ORI, Hauser and Harvard were not and are not at liberty to discuss the affair: confidentiality is a well-justified, codified policy of Harvard (8) (indeed, the majority of universities) and the ORI (6).
Such are the publicly available facts; they are limited and uninformative, as expected for any such investigation, not just Harvard’s and ORI’s. The latter typically accepts the findings of the former, ultimately issuing equally limited findings (6, 8). Despite these limits, though, a feeding frenzy began among the headline-hungry media as well as scholars opposed to Hauser’s research program, with conjectures as to the nature of the misconducts (9), insinuations that all of Hauser’s work is (and always has been) dubious (10, 11), and demands that the confidentiality of the federal investigation—and thereby Hauser’s presumption of innocence (which he does retain in an ongoing investigation)—be breached (12). This ungrounded gossip and sensationalism defiled common sense and common decency, distracting the public—nonscientists and scientists—from the scandal of greater importance: the general nature of investigation into misconduct. It has a distinctive ring of McCarthyism, and all the toxic implications of that witch-hunting era. This is an issue larger than Marc Hauser, with implications for the very integrity of science—the central issue of this essay. Thus it is necessary to distinguish the facts of the case from the process of its prosecution (the manner in which the facts are procured); the issue of whether the former, particular to a case, are condemnatory or exonerative is not so important as whether the latter, general to all cases, is just. Facts are adjudged only within the procedural framework of a case such that if the process is unjust, then any judgment on the facts is untrustworthy and establishes a dangerous precedent to which any of us could be subject.
Distinguishing the particulars of the Hauser case from its general implications is exceedingly difficult, considering the unjust and fallacious presumptions of certain academics opposed to Hauser’s research program that not only is Hauser guilty in this particular affair, but that now the totality of his work—and to some extent that of the field in toto (13)—is suspect (10, 11, 14). Such claims are patently absurd and should offend all researchers by reducing science from an (imperfect) forum for rational inquiry and assessment to a kangaroo court. If Hauser, or—to make clear the general point—anyone with a long research career, were a fundamentally flawed and deceptive scientist, their work would not have been so numerously, diversely, and robustly replicated in so many permutations of species and methodologies both by his lab and others. In Hauser’s particular case, a scan of Wikipedia (15) or a search via any publication engine (e.g., Google Scholar) quickly reveals fifty or more of the many studies that have replicated his groundbreaking originals in hemispheric specialization, linguistic function, comprehension of goal-directed action, tool use, and folk physics. Replication is the (difficult) stuff of good science (16). Obviously, replicating an experiment does not prove that the original was not fraudulent, but it is certainly implausible for a multiplicity of fraudulent originals from a multiplicity of domains to be so multiply replicated with multiple methods in multiple species. Such a record is significant, for replications of studies with nonhuman primates are, in general, infrequently conducted and nontrivial: sample sizes are small so that the effects of individual variability are considerable; few labs have the same species so that attempted replications are in different species testing for the same capacity; and whether the animals are captive or wild, the peculiarities of the environment can be difficult to duplicate (something true with animal behavior generally).
The replication record furthermore makes manifest the general point that a principal investigator’s generation of students go on not only to continue the line of research, but to extend it in important ways. Thus, productive scientists often generate a lineage of equally productive students. These students often enter the lab and work on projects that have previously been carried out by the principal investigator and his previous students. If the earlier work was fallacious, the lab-internal replications fail. If the principal investigator was running a shoddy operation, it is unlikely that his students would go on to productive careers. Hauser has not only trained some of the most distinguished figures in the field of cognitive evolution, but many of his students have gone on to replicate and extend a number of his fundamental findings. In fact, we know of no case where his students have failed in these replications. The latter is certainly something that happens in science, and this is one of its healthy attributes. To put it simply, a panoply of Hauser’s students are now prominent in their own rights; would they be so esteemed and able to replicate and extend their mentor’s work if they had trained under a regime of fabrication, falsification, and plagiarism? Is it rational to believe that Hauser is the godfather of a secret international cartel for scientific fraud? That we can even pose such questions shows just how dysfunctional a misconduct investigation can be, and how toxic to science the media (and the media-hungry) can become. None of this is to say that Hauser did not make mistakes, as he himself admitted to the public (17).
To demonstrate how any scientist accused of misconduct deserves to be treated by his peers, let us now briefly consider the facts of the three problems in Hauser’s published work so that we can build the more general case. The papers were published in Proceedings of the Royal Society B (18), Science (19), and Cognition (20). It is curious that Hauser is accused of misconduct in the sense of fabrication or falsification given that, according to Harvard (4), “the experiments reported [by the Harvard investigating committee] were designed and conducted” (emphasis added); i.e., the experiments were not fabricated. And it is curious that Hauser could be “solely responsible” for any misconduct given that for the first and second studies, it was not he who collected or coded the data. What does this say about the nature of responsibility? These curiosities merit further consideration. Of course, Hauser, as principal investigator (and lead author on the first paper, coauthor on the second and third), is ultimately responsible for the work of his lab, but it would be impossible (practically) and overbearing (for his students who are maturing into independent researchers) for him to micromanage all experiments, verifying all details. This is true—and rightfully so—for all labs, not just Hauser’s.
The first and second studies have been replicated to correct what those journals defined as inadvertent failures to archive some of the original field notes and videos (21, 22). These were problems in data retention, not instances of falsification or fabrication or plagiarism. It is thus quite curious that Harvard cited these studies as instances of misconduct. This just shows how arbitrary and capricious charges of misconduct can be. The third study was retracted because “the data do not support the reported findings” (23). (However, the findings are consistent with the data from numerous replications and related studies by Hauser and his colleagues in addition to independent work (even on different species) by labs unaffiliated with Hauser (24-26).) The retraction statement is vague, inviting invidious speculations (9) and plausible exonerative inferences (27) alike. Presumably, had Hauser been solely responsible for misconduct in this instance, the retraction would have stated so; it did not. But most informative and credible is the testimony of Professor Bennett G. Galef, who, at the request of Hauser and his attorneys, reviewed Harvard’s evidence for misconduct in the three published cases (28), concluding thus: “In my opinion there is nothing in the charges and data relating to the three published articles to show Hauser guilty of anything. Rather, what I saw on numerous occasions in numerous ways was Harvard violating what I take to be elementary principles of natural justice” (29); principles to which all scientists—all people—are entitled. Still the nature and scope of the facts are limited so any judgment must be provisional. More important, as we have argued, is the nature and scope of the investigatory process which, to reemphasize, impinges upon—and infringes upon—the freedom of science.
To conclude, we do concur with the critics that the Hauser affair is dangerous for science (10, 11, 14). However, unlike the critics, we argue that the danger is posed not by Marc Hauser, but rather by the process of his prosecution. As scientists, we all ought to be scared by the idea of an inquisitorial method that abolishes the scientific method by prying into unpublished work, feeds a media frenzy (thus luring in federal prosecution), and overlooks the hypocrisy of the critics who fail to look at their own vulnerabilities and how they run their own labs. But we cannot be scared out of our wits, for we will need all our wits about us to face this clear and present danger.
References
1. N. Wade, “In Harvard lab inquiry, a raid and 3-year wait” (The New York Times, 13 August 2010).
2. T. Bartlett, “Document sheds light on investigation at Harvard” (The Chronicle of Higher Education, 19 August 2010).
3. G. Miller, “Hausergate: Scientific misconduct and what we know and don’t know” (Science Insider, 25 August 2010).
4. M.D. Smith, letter from the dean of arts and sciences to the Harvard faculty, 20 August 2010.
5. J.S. Mill, On Liberty and Other Writings (Cambridge: University of Cambridge Press, 1859/1989), pp. 20.
6. http://ori.hhs.gov/policies/fed_research_misconduct.shtml
7. S. Titus, X. Bosch, Tie funding to research integrity. Nature 466, 436-437 (2010).
8. http://www.fas.harvard.edu/~research/greybook/misconduct.html#f0
9. G. Altmann in “Fabrication plausible, journal editor believes,” C.Y. Johnson (The Boston Globe, 28 August 2010).
10. F.B.M. de Waal in “Harvard says Marc Hauser guilty of science misconduct,” D. Vergano (USA Today, 20 August 2010).
11. G. Gallup in “Author on leave after Harvard inquiry,” C.Y. Johnson (The Boston Globe, 10 August).
12. M. Tomasello in “Misconduct probe in Harvard animal morality lab,” P. Aldhous (New Scientist, 18 August 2010).
13. F.B.M. de Waal in H. Ledford, Harvard probe kept under wraps. Nature 466, 908-909 (2010).
14. M. Tomasello, R. Seyfarth in “Inquiry on Harvard lab threatens ripple effect,” N. Wade (The New York Times, 12 August 2010).
15. http://en.wikipedia.org/wiki/Marc_hauser (accessed on 27 June 2011)
16. C. Zimmer, “It’s science, but not necessarily right” (The New York Times, 25 June 2011).
17. M.D. Hauser in “Harvard psychology professor Hauser’s statement” (The Boston Globe, 20 August 2010).
18. M.D. Hauser, D.D. Glynn, J.N. Wood, Rhesus monkeys correctly read the goal-relevant gestures of a human agent. Proceedings of the Royal Society B 274, 1913-1918 (2007).
19. J.N. Wood, D.D. Glynn, B.C. Phillips, M.D. Hauser, The perception of rational, goal-directed action in nonhuman primates. Science 317, 1402-1405 (2007).
20. M.D. Hauser, D. Weiss, G. Marcus, Rule learning by cotton-top tamarins. Cognition 86, B15-B22 (2002).
21. M.D. Hauser, J.N. Wood, Replication of ‘Rhesus monkeys correctly read the goal-relevant gestures of a human agent.’ Proceedings of the Royal Society B 278, 158-159 (2011).
22. J.N. Wood, M.D. Hauser, Replication of ‘The perception of rational, goal-directed action in nonhuman primates.’ Science 332, 537 (2011).
23. M.D. Hauser, D. Weiss, G. Marcus, Retracted: ‘Rule learning by cotton-top tamarins.’ Cognition 117, 106 (2010).
24. A.D. Endress, S. Carden, E. Versace, M.D. Hauser, The apes’ edge: Positional learning in chimpanzees and humans. Animal Cognition 13, 483-495 (2010).
25. M.D. Hauser, D.D. Glynn, Can free-ranging rhesus monkeys (Macaca mulatta) extract artificially created rules comprised of natural vocalizations? Journal of Comparative Psychology 123, 161-167 (2009).
26. R.A. Murphy, E. Mondragón, V.A. Murphy, Rule learning by rats. Science 319, 1849-1851 (2008).
27. N. Wade, “Harvard case against Marc Hauser is hard to define” (The New York Times, 25 October 2010).
28. N. Wade, “A journal’s statement may aid a Harvard researcher accused of misconduct” (The New York Times, 25 April 2011).
29. B.G. Galef, personal communication.
