Studies in History and Philosophy of Biological and Biomedical Sciences 44 (2013) 60–70

Contents lists available at SciVerse ScienceDirect

Studies in History and Philosophy of Biological and

Biomedical Sciences
journal homepage: www.elsevier.com/locate/shpsc

Science, truth, and forensic cultures: The exceptional legal status of DNA evidence
Michael Lynch
Department of Science & Technology Studies, Cornell University, 302 Rockefeller Hall, Ithaca, NY 14853, USA

a r t i c l e i n f o a b s t r a c t

Article history: Many epistemological terms, such as investigation, inquiry, argument, evidence, and fact were estab-
Available online 30 October 2012 lished in law well before being associated with science. However, while legal proof remained qualified
by standards of ‘moral certainty’, scientific proof attained a reputation for objectivity. Although most
Keywords: forms of legal evidence (including expert evidence) continue to be treated as fallible ‘opinions’ rather
DNA profiling than objective ‘facts’, forensic DNA evidence increasingly is being granted an exceptional factual status.
Fingerprinting It did not always enjoy such status. Two decades ago, the scientific status of forensic DNA evidence was
Expert evidence
challenged in the scientific literature and in courts of law, but by the late 1990s it was being granted
Forensic science
Objectivity
exceptional legal status. This paper reviews the ascendancy of DNA profiling, and argues that its
Moral certainty widely-heralded objective status is bound up with systems of administrative accountability. The ‘admin-
istrative objectivity’ of DNA evidence rests upon observable and reportable bureaucratic rules, records,
recording devices, protocols, and architectural arrangements. By highlighting administrative sources of
objectivity, this paper suggests that DNA evidence remains bound within the context of ordinary organ-
isational and practical routines, and is not a transcendent source of ‘truth’ in the criminal justice system.
Ó 2012 Published by Elsevier Ltd.

When citing this paper, please use the full journal title Studies in History and Philosophy of Biological and Biomedical Sciences

1. Introduction one is inclined to accredit it with ‘scientific’ status, the evidence

adduced through the practice is no less bound up in forensic cul-
This paper focuses on the exceptional legal status currently ture than any other form of criminal evidence.
granted to DNA profile evidence. DNA profiling is the ascendant Strong justifications currently are given for the overriding
‘gold standard’ in criminal forensics, though not very long ago fin- weight assigned to DNA evidence: many of the constituent tech-
gerprint identification held the title. The convictions (in both the niques derive from molecular biology and are widely established
epistemological and legal senses of the word) associated with con- in numerous fields of research and diagnostic practice; early prob-
fessions, eyewitness testimony, and many other forms of evidence lems with forensic uses of the techniques were aired in admissibil-
are now held to be fallible and ‘subjective’ in comparison with DNA ity hearings, debated in the scientific literature, and subjected to
evidence. When compared with what Simon Cole in his article in numerous inquiries and audits; and refinements in technique
this special section calls ‘research science’, forensic science is often and improvements in methods of probabilistic estimation have
granted separate and unequal status as a knowledge-producing remedied the early problems. In brief, it is often said that DNA pro-
culture. However, DNA profiling is often treated as an exceptional filing has ‘passed the test’, both in science and in law, unlike many
technique that has navigated the imaginary gulf between forensic other forms of ordinary and technical evidence. It is now common
and scientific culture, and has become so reliable that it can be to use the shorthand expression ‘DNA’ as a reference to the various
used as a benchmark for evaluating all other forensic technologies. techniques of DNA profiling, and to treat ‘it’ as both the source and
Like Paul Roberts in his contribution to this section, I argue that object of an extraordinary, and even absolute, degree of certainty
such comparisons misleadingly associate ‘science’ with certainty in criminal law. I shall argue, however, that granting such excep-
and ‘forensics’ with practical contingency. DNA profiling remains tional status to DNA evidence ignores the fact that it is adminis-
an integral part of criminal investigations, and whether or not tered alongside other types of evidence in a criminal justice

E-mail address: mel27@cornell.edu

1369-8486/$ - see front matter Ó 2012 Published by Elsevier Ltd.

http://dx.doi.org/10.1016/j.shpsc.2012.09.008
 M. Lynch / Studies in History and Philosophy of Biological and Biomedical Sciences 44 (2013) 60–70 61

system. Moreover, its probative value depends upon other forms of an environment where participants engaged in creating or
evidence, including testimony about how particular evidence sam- determining the ‘truth’ of something by a set of site-specific
ples were handled, analysed, and interpreted. The practical and rules, we can see that juries, judges, witnesses and counsel par-
administrative contingencies associated with ‘DNA’s’ probative va- ticipated in a process designed to produce ‘morally certain’
lue should not be a revelation to anyone who is familiar with the determinations of ‘matters of fact’.3
way DNA evidence is collected, analysed, and presented in criminal
Contrary to arguments that emphasise the religious or eco-
trials, but, I shall argue, it is crucial to include such contingencies in
nomic roots of science, Shapiro points to legal procedures as pro-
any characterisation of ‘DNA’ and what ‘it proves’.
viding a model for Robert Boyle’s effort to solicit and interrogate
the ‘testimony of nature’.4 This is more than a matter of ‘values’
2. Moral and mathematical certainty being carried over from one institution to another, as it involves
common administrative arrangements designed to produce a pecu-
Law and science share a common vocabulary that features liar mode of objectivity later associated with the ideal bureaucracy:
terms such as ‘investigation’, ‘inquiry’, ‘argument’, ‘evidence’, ‘fact’, case-by-case judgments by designated officials who justify their
‘law’, ‘test’, and ‘proof’. Although these terms currently have dis- decisions as disinterested applications of impersonal rules. However,
tinct (and even divergent) uses in law and science, they have com- Shapiro also observes that by the eighteenth century, a clear distinc-
mon seventeenth-century origins and they continue to be tion was drawn between the ‘moral certainty’ proper to the court-
associated, sometimes in a confusing way, in cases in which law room, and the ‘mathematical certainty’ to which science aspired:
bears upon science and science upon law. Historian Barbara Shap- ‘Jury verdicts could never be proved with the certitude that early
iro points out that in medieval canon law, the concept of fact ‘re- modern persons associated with the terms ‘‘science’’ or ‘‘mathemat-
ferred to a human event or act that had taken place in the past. ical certainty’’. The language employed to indicate the highest kind
In the legal context it referred to a past act, for instance a particular of probability available in ‘‘matters of fact’’ was ‘‘a satisfied con-
homicide or theft, that could not be observed first-hand in the science’’, ‘‘moral certainty’’ and later ‘‘beyond reasonable doubt’’’.5
courtroom. Fact did not refer to natural events or natural phenom- Although moral certainty was deemed the highest degree of
ena’. With the decline of ‘irrational proofs’ based on trial by ordeal, certainty attainable in the context of a jury trial, the contrast with
testimony and credibility became crucial: mathematical certainty implied that it fell short of the epistemic
. . . ‘facts’ might be fully proved either by confession of the standard ascribed to mathematical and experimental proof. As
accused or the testimony of two credible witnesses. Credibility Shapin and Schaffer point out, however, the ideal of mathematical
was determined by criteria such as gender, age, social status certainty was upheld by critics (particularly Thomas Hobbes), who
and education. The nature of the witness’ friends and associates denigrated the epistemic status of the ‘matters of fact’ arising from
and his life style were relevant. Determinations of both fact and Robert Boyle’s pneumatic experiments. In line with that ideal, a
law were in the hands of highly educated, professional judges, competent witness to a mathematical or experimental demonstra-
who employed these criteria to evaluate credibility. The testi- tion should be able directly to follow the steps in a proof, but in the
mony of one credible witness, which constituted a ‘half proof’, case of testimony (whether by a witness or an experimental
was sufficient to trigger judicial torture to elicit a confession arrangement) proof is indirect and requires fallible practical judg-
constituting ‘full proof’.1 ment. In John Locke’s famous formulation, an impartial evaluation
of testimony takes into account ‘the number of witnesses, their
Shapiro goes on to say that during the early modern period, the integrity, their skill at presenting evidence, and its agreement with
roles of jurors and witnesses became more prominent for furnish- the circumstances, and . . . the presence or absence of contrary tes-
ing and evaluating testimony. In the late middle ages, the two roles timony.’6 These proxies for truth, though perhaps reliable as bases
were combined, as jurors were selected from the community in for credibility judgments, were explicitly recognised in legal proce-
which the alleged crime or dispute took place and they were ex- dures as fallible and open to challenge. Shapin argues that Boyle’s
pected to draw upon their first-hand knowledge of the parties experimental programme was ‘probabilistic’ (in a non-mathematical
and particulars. Starting in the sixteenth century, however, courts sense of the word), and that the matters of fact produced through
gradually separated the witness and juror functions in the interest that programme were always more or less true, and subject to
of impartiality. Shapiro speculates that this change was connected down-to-earth considerations akin to those listed by Locke.7 Though
with the massive population movements and concentration in never granted absolute certainty, experimental facts gradually ac-
industrial centers, which disrupted the continuities of habitation quired a degree of credibility that distinguished the epistemic status
and local knowledge associated with traditional village life.2 The of ‘scientific’ facts from the ‘moral’ certainty associated with legal
selection of uninvolved jurors separated the role of the witness from determinations.
the jury’s evaluative function, so that the credibility and plausibility The contrast with mathematical certainty, and later ‘scientific’
of testimony, as judged by a disinterested party, was now of para- fact, did not necessarily imply that moral certainty was an inade-
mount importance. Alluding to Shapin and Schaffer’s genealogy of quate standard for deciding legal facts. Indeed, in Anglo-American
‘matters of fact’ in the Royal Society’s experimental programme, jurisprudence, starting in the mid-nineteenth century, ‘beyond a
Shapiro points to a shift in legal perspective towards facts that coin- reasonable doubt and to a moral certainty’ was the recommended
cided with the administrative reforms that positioned the jury as a standard for convicting a defendant in a criminal trial. It was rec-
disinterested ‘third party’ in the courtroom. ommended because it did not demand such a high a standard of
If we consider the courtroom—as others have considered the proof that conviction would become practically impossible. ‘Mor-
scientific experiment—as a site of knowledge-making, that is, ally certain beliefs could not be proven beyond all doubt but they

1
Shapiro (2002), p. 245.
2
Shapin & Schaffer (1985), Shapiro (1986), p. 155.
3
Shapiro (2002), p. 247.
4
See also Shapin (1994, 1995).
5
Shapiro (2002), p. 248; see also Shapiro (1986, 1991).
6
Locke (1995[1690]), p. 557.
7
Shapin (1994).
62 M. Lynch / Studies in History and Philosophy of Biological and Biomedical Sciences 44 (2013) 60–70

were nonetheless firm and settled truths, supported by multiple clude institutional arrangements designed to ensure disinterested-
lines of evidence and testimony’.8 ness and impartiality, and both emphasise administrative as much
In the late twentieth century, U.S. courts deemed ‘moral cer- as (or even more than) personal virtues, though the latter also are
tainty’ to be a quaint and misleading formulation. Part of the con- important, as Shapin chronicles.14 Both Weber’s ideal type of
cern was that contemporary jurors would understand the word bureaucracy and Merton’s idealised scheme of scientific norms were
‘moral’ as a vernacular expression referring to personal and emo- criticised for their lack of any clear relationship to routine practices
tional opinion, rather than conviction based on a disinterested and negotiations, whether with clients in an office or recalcitrant
evaluation of evidence. Consequently, the phrase ‘to a moral cer- materials in a laboratory.15 Indeed, the ‘new’ sociology of science
tainty’ was removed from judicial instructions to juries about the beginning in the 1970s cut its teeth against Merton’s norms by pur-
appropriate standard of proof in criminal trials, leaving ‘beyond a suing historical and sociological case studies on the ‘construction’ of
reasonable doubt’ as the explicit standard. Laudan argues that de- scientific facts.16 In line with sceptical philosophical conceptions of
spite judicial efforts to avoid such an outcome, ‘beyond a reason- scientific knowledge, the picture of experimental science developed
able doubt’ subsequently became a degraded standard, often in constructionist studies breaks down the distinction between mor-
understood as a reference to ‘jurors’ subjective states of mind’ al and mathematical certainty. According to a ‘finitist’ philosophy of
rather than to the probative value of the evidence against the science, no finite array of data supports one and only one theoretical
defendant.9 explanation; indefinitely many interpretations are always possible.
Aside from how the words ‘moral certainty’ are, or were, under- Consequently, scientists must make practical judgments about when
stood, what Shapiro describes is more than an epistemic standard. to quit collecting data and to declare a conjecture confirmed or not.17
As she states in the above passage, moral certainty was bound up Empirical findings are thus tentative, subject to revision, and never
in an institutional ‘process designed to produce ‘‘morally certain’’ grounded in absolutes.18 In other words, the certainties involved
determinations of ‘‘matters of fact’’.’ This process included court- are moral certainties. Mathematical certainty becomes an impossible
room protocols for soliciting testimony, cross-examining wit- ideal, and testimony and credibility judgments are no less a part of
nesses, selecting jury members, and reaching verdicts through the practice of science than they are of law.19
deliberation.
Because it was believed that biased judges, juries and witnesses 3. Scientific evidence in a legal context
contaminated legal fact-finding, institutional safeguards, how-
ever imperfect, were developed to prevent interested testimony The expression ‘to a moral certainty’ may have fallen out of fa-
and judgment. The norm of impartiality was also adopted by vour in the U.S. courts, but the distinction remains very much alive
the Royal Society, and members often proclaimed their lack of between ‘scientific’ evidence of a kind that is deemed beyond dis-
interest and bias. Boyle claimed to cultivate ‘chemistry with a pute and all other forms of evidence (including technical, forensic,
disinterested mind’, without caring for ‘personal advantages’.10 and some weaker modes of ‘scientific’ evidence). There is a key dif-
ference between this distinction and the traditional one between
Readers familiar with the sociology of science are likely to asso- moral and mathematical certainty. The standard of mathematical
ciate ‘the norm of impartiality’ with Robert K. Merton’s norms of certainty was assumed to be too stringent for practical use in the
science: communalism, universalism, disinterestedness, and courtroom, so that moral certainty was, by default, the best that
organised scepticism.11 Indeed, Merton derived his norms through could be attained. In the present-day courtroom, the certainty as-
an analysis of edifying statements such as those Shapiro quotes from cribed to exceptional forms of ‘scientific’ evidence is very much
Boyle. Merton famously argued that these norms were a transmuta- in play within the trial court. This distinction between standards
tion of the Protestant ethic—a variant of worldly asceticism that Max of certainty is not formally codified in legal procedure, but it is ana-
Weber identified with the ‘spirit of capitalism’. In an argument that lytically associated with particular legal efforts to achieve practical
was parasitic on Weber’s thesis, Merton alleged that the complex of demarcations.
values associated with the Protestant ethic motivated the gentlemen Many philosophers of science abandoned the effort to demar-
of the Royal Society to direct their energies to a selfless accumulation cate science from non-science in the wake of Popper’s influential,
and reinvestment of symbolic capital.12 but ultimately unsuccessful, effort to establish falsifiability as a
Although Merton presented the complex of norms as being dis- demarcation criterion.20 This has not deterred judges from invoking
tinctive of the social institution of science, there is substantial Popper’s criterion (or others such as ‘methodological naturalism’) to
overlap with Weber’s ideal type of rational bureaucracy.13 Both in- define what counts as science as opposed to religion (McLean v.

8
Laudan (2003), p. 297.
9
Ibid., p. 300, p. 330.
10
Shapiro (2002), p. 258.
11
Merton (1973[1942]).
12
Weber (1976), Merton (1970).
13
Weber (1968).
14
Shapin (1994).
15
March & Simon (1993[1958]) distinguished their theory of organisations from Weber’s model of bureaucracy. The many critiques of Merton’s norms include Barnes & Dolby
(1970), Mitroff (1974), and Mulkay (1976). In his article in this special section Cole (2013) acknowledges such criticisms, but argues that Merton’s norms provide a rough basis for
contrasting the ‘epistemic culture’ of research science to that of forensic science.
16
Bloor (1976) set out the new ‘strong programme’ in the sociology of science, and Latour & Woolgar (1979) developed the theme of the ‘construction of scientific facts’ in their
ethnographic study.
17
Collins (1985).
18
Finitism is a term used by Barnes, Bloor, & Henry (1996) to summarise lines of argument in philosophy of science that emphasise the insufficiency of finite arrays of empirical
data (or, in the case of mathematics, examples) for proving symbolic generalisations beyond all possibility of doubt. Such scepticism about certainty in science runs through a line
of philosophy of science, including Popper’s critique of induction, Duhem and Quine’s underdetermination thesis, and Hanson’s thesis about the theory ladenness of observation.
19
Shapin (1995).
20
Popper (1959, 1963).
 M. Lynch / Studies in History and Philosophy of Biological and Biomedical Sciences 44 (2013) 60–70 63

Arkansas, 1982; Kitzmiller v. Dover Area School District, 2005), or to An example was the case in Pennsylvania of Bruce Godschalk,
define what counts as a reliable foundation of admissible expert evi- who in 1987 had been sentenced to a lengthy prison term for rap-
dence (Daubert v. Merrell Dow Pharmaceuticals, Inc., 1995).21 From a ing two women in separate incidents. When his case was re-
philosophical point of view, there may be no stable criteria for dis- opened 15 years later, the defence experts presented DNA profiles
tinguishing science from non-science, but from a sociological or his- developed from the semen evidence recovered from the two vic-
torical vantage point, legal (and other) efforts to demarcate tims. Godschalk’s defence argued that the profiles matched each
(genuine) science from non-science have a constitutive role in the other, but did not match their client’s profile. Though he did not
establishment and reification of consequential institutional bound- present DNA evidence contradicting such results, the prosecutor
aries.22 A similar stance can be taken toward the degree of certainty who originally had handled the case refused to release Godschalk
ascribed to ‘scientific’ evidence, as compared with other forms of le- from prison, arguing that the DNA tests were less convincing than
gal evidence. Though problematic for contemporary science studies, the confession Godschalk made to the police and later recanted.
the distinction has a place in legal institutions, and its use has organ- The prosecutor was quoted as saying: ‘I have no scientific basis, I
isational consequences. For sociologists of science-in-law, the dis- know because I trust my detective and my tape-recorded confes-
tinction makes up a relevant and important topic even if it is a sion. Therefore the results must be flawed until someone proves
dubious conceptual resource.23 to me otherwise’.29 However, the Innocence Project lawyers and
the expert witnesses they had contacted convinced the court that
4. DNA exceptionalism they had furnished such proof. Like Lloyd, they likened a DNA profile
to a signature, and argued that matching profiles can only come from
Courts in the U.S. have placed ‘DNA’ in an exceptional epistemic one person (or at the most a very few individuals) in the world, and
space, if one is to believe the chorus of testimonials in recent years that in the Godschalk case the chances of laboratory error were
voiced by prosecutors, defence attorneys, judges, legal scholars, ‘nonexistent’.30
and persons who have been exonerated with DNA evidence.24 Even
convicted felons serving jail terms testify to the power of ‘DNA’.25 It 5. The ascendency of DNA ‘fingerprinting’
is described as ‘infallible’, as furnishing ‘irrefutable proof’, and as a
‘truth machine’ or ‘revelation machine’.26 Perhaps the most widely The certain, and even transcendent, epistemic status granted to
circulated quote (which may have been a misquote) is attributed DNA evidence is quite remarkable, though not unprecedented.
to former prison inmate Eddie Joe Lloyd: ‘DNA–deoxyribonucleic Well before DNA profile evidence was admitted into criminal trials
acid–is God’s signature. God’s signature is never a forgery and his in the U.K., U.S., and many other nations, fingerprint evidence had
checks don’t bounce!’27 Lloyd had served 17 years in prison on a attained ‘factual’ status in criminal trials. Jennifer Mnookin distin-
rape conviction before DNA testing of the original evidence was guishes the ‘factual’ status assigned to fingerprint evidence from
instrumental for exonerating him. Lloyd, who unfortunately died the evidentiary status of expert ‘opinion’.31 This status was, so to
soon after being freed from prison, was one of the, by now, more speak, de facto rather than formally codified or based on any specific
than 250 prisoners whose original verdicts were overturned through legal ‘test’. Mnookin quotes from the judge in the 1917 trial People v.
DNA testing or retesting of blood, semen, or other bodily evidence Sallow, ‘The witness does not testify. The physical facts speak for
that had been collected during the original investigation and (fortu- themselves; no fears, no hopes, no will of the prisoner to falsify or
nate for the prisoner) retained following the conviction.28 In many to exaggerate could produce or create a resemblance of her finger
cases, the convictions were based on eyewitness testimony, confes- prints or change them in one line, and therefore there is no danger
sions, and other forms of forensic evidence. In some cases, retesting of error being committed or untruth told.’32 She quotes another
with newer DNA profile methods overturned convictions based on judge in a 1930 case, who asserts, ‘fingerprinting is based on the
older, less discriminating, DNA tests. One of the challenges faced law of nature, or upon a universally recognized physical fact.’33
by the Innocence Project was to convince judges and district attor- The judge’s statement is as clear an invocation as one could
neys to allow cases to be reopened, often years after all appeals want of an idea (or ideal) of objective evidence that derives directly
had been exhausted. Even when such cases were reopened, and de- from nature, with human mediation being too trivial to mention. In
spite the fact that DNA evidence was, more often than not in criminal this case, the judge discounts the possibility that ‘the prisoner’
cases, a prosecutorial tool, some prosecutors resisted exonerating could influence her fingerprints, rather than that an observer
evidence. would willfully or inadvertently affect the evidence. In Goffman’s

21
For critical discussions of these cases see Solomon & Hackett (1996), Jasanoff (1995), Edmond (2002), Edmond & Mercer (1997), Haack (2005), Caudill & Redding (2000); and
Caudill & LaRue (2006).
22
Gieryn (1983, 1999).
23
The topic-resource distinction derives from ethnomethodology (Garfinkel & Sacks, 1970; Zimmerman & Pollner, 1970). The idea was that many of the concepts and analytical
activities that sociologists deploy as methodological resources are themselves constitutive features of the social organisational affairs that sociologists attempt to describe.
Ethnomethodologists consequently attempted to make a topic of ‘methodology’, rather than to use methodology as a technical basis for claiming scientific authority over and
against the ordinary practices at large in the societies studied.
24
Lynch, Cole, McNally, & Jordan (2008), p. 261ff.; Aronson & Cole (2009).
25
Prainsack & Kitzberger (2009).
26
See Berger (2004), Scheck, Neufeld, & Dwyer (2000); and Findley (2002) for testimonies to the extraordinary power of DNA testing. Prainsack & Kitzberger (2009) document
the uncritical acceptance by Austrian prison inmates that it is pointless to contest DNA evidence.
27
Different versions of this quotation were presented in press reports and on various websites. The phrase ‘. . . his checks don’t bounce’ is doubtful, and a more accurate
version may be: ‘That’s God’s signature. God’s signature is never a forgery’ (Wilgoren, 2002). However formulated, this quotation, as an exemplar of stories of DNA-
exonerations, has a mythic quality that mirrors stories of DNA-convictions, such as the one about the conviction of Steve Wright discussed in Prainsack & Toom’s article in this
special section (2013).
28
Scheck et al. (2000).
29
Rimer (2002).
30
Ibid. For a discussion of continuing efforts by prosecutors to resist exonerations through DNA testing, see Martin (2011).
31
Mnookin (2001), p. 30; also see Cole (2001).
32
Ibid., p. 27, quoting from (People v. Sallow, 1917).
33
Dissent in State v. Steffan, Iowa 1930, quoted in Mnookin (2001), p. 30. Also see People v. Jennings (1911).
64 M. Lynch / Studies in History and Philosophy of Biological and Biomedical Sciences 44 (2013) 60–70

terms, the prisoner ‘gives off’ the incriminating signs, despite any proved that they were not (and never were) competent
effort to conceal them, and an unmentioned analyst simply reads examiners.35
the signs.34 The logic of criminal evidence supplements the classic In the mid-1980s, when DNA profiling techniques were first
epistemological mediation between observer and observed with a developed for use in criminal investigations, fingerprint examina-
more primary mediation between criminal and evidence: the possi- tion was so well established that it provided a model (and a name)
bility of inadvertent deception preoccupies analysis of the former, for what was popularly dubbed ‘DNA fingerprinting’. There was a
while the possibility of deliberate deception preoccupies analysis basis for the analogy: both techniques were used as tools in crim-
of the latter. Nature’s agency short circuits human mediation, and inal investigations, and both were used to identify suspects on the
by doing so defeats the criminal’s deliberate efforts to control the basis of trace evidence found at crime scenes. In addition, the ear-
evidence and then speaks directly, with unmistakable clarity, to liest technique developed for criminal investigations in the U.K. in-
the observer. volved the use of radioactive markers that bonded to particular
Mnookin draws upon Simon Cole’s excellent historical study of DNA sequences at an indefinitely large number of genomic loci.
fingerprinting, and notes that courts and legal commentators con- This ‘multi-locus probe’ technique yielded a complex pattern that
flated two, logically distinct, aspects of fingerprint evidence. The was also likened to a bar code. Given the complexity of the resul-
first was the dogma that fingerprints (or, more generally, dermal tant pattern, it was deemed virtually impossible to find coinci-
ridge patterns on the fingertips and elsewhere on the hands and dence matches between profiles developed from the blood or
feet) were individually unique identifiers that remained stable other DNA-containing bodily material from different individuals.
throughout a person’s lifetime. The second was that professional A different ‘single-locus probe’ technique was initially adopted in
fingerprint examiners were capable of making error-free judg- the U.S., and shortly thereafter it supplanted the multi-locus probe
ments about whether or not two fingerprints matched. In many technique in the U.K.. This technique used a finite number of mark-
criminal cases, the task for the examiner was to determine if a ‘la- ers, each of which would ‘label’ a single locus. The advantage of the
tent’ fingerprint lifted from a surface at a crime scene (a window- single-locus probe technique was that the resulting patterns were
sill, jewelry box, etc.), and circumstantially associated with the more amenable to controlled use and probability calculation. Ini-
perpetrator’s imagined movements, matched a print on file or ta- tially (in the late 1980s and early 1990s), the number of markers
ken under controlled conditions from a suspect in custody. ‘Latent’ was limited to four or six, and the random match probabilities cal-
prints tend to be partial, often faint or smudged and superimposed culated from matching profiles were on the order of one in hun-
on the texture of the surface from which they were lifted. The con- dreds of thousands or a few million. These were impressive
tinuity in the quality of latent fingerprints—ranging from com- figures when compared with, say, the probabilities associated with
pletely indecipherable smudges to clear and readable traces of matching blood types, but they left open the possibility that more
dermal ridge patterns—presented a potential problem for examin- than one individual in a large urban or national population would
ers that was ‘solved’ with a conservative strategy. Examiners were match a profile developed from criminal evidence. Moreover, the
instructed to refrain from declaring matches, unless the prints methods for calculating random match probabilities involved sta-
were of sufficient quality to make a determination and a sufficient tistical and population genetic assumptions that were heavily crit-
number of ‘points’ (codified ridge characteristics, such as whorls icised in scientific publications and legal hearings.36
and bifurcations) were found to match in the compared prints Fingerprint examiners faced no such complications: an exam-
without any mismatched points. Scotland Yard used a 16-point iner simply could declare that the evidence matched, did not
threshold, whereas other national or regional forces used lower match, or was inconclusive. In addition, unlike the routines of fin-
or higher numbers of points. Point systems have been abandoned gerprint collection and analysis, forensic DNA analysis involved a
by most forensic organisations, as they do not translate into spe- more complicated, esoteric, and delicate set of practical operations.
cific probability figures. Although a given number of points de- A litany of errors and possibilities for error began to accumulate, as
noted a presumptive threshold of certainty, this threshold varied defence attorneys began to learn of the technique and its (mis)uses
arbitrarily between different national systems. through consultations with molecular biologists, population genet-
In the U.S. until recently, a circular logic was used to preserve icists, and statisticians.37 In addition to citing laboratory errors, crit-
the claim that fingerprint examination was error free. By the ics of the technique identified ‘pre-analytic’ errors made by police
1930s, fingerprint evidence was widely accepted by the courts, detectives and others who collected evidence, put labels on the evi-
and judges, lawyers, and defendants assumed that fingerprint dence, transported it to labs, and placed it in storage. Although DNA
matches were certain and virtually unassailable as individual iden- evidence proved persuasive in many criminal trials, some key court-
tifiers. Fingerprint evidence was rarely contested, but occasionally room challenges occurred in the U.S., the U.K., Australia, and else-
examiners’ declarations were challenged during trials and deter- where. The analogy with fingerprinting no longer seemed so
mined to be erroneous. The official policy of the fingerprint exam- salient, and forensic scientists began to discourage the use of the
iners’ professional association in such cases was to suspend the popular name ‘DNA fingerprinting’. At the time (the early 1990s), la-
accreditation of any examiners held responsible for such errors. tent fingerprinting identification appeared simpler and more trans-
An error was defined as a personal mistake, and the source of the parent: presenting and interpreting fingerprint evidence seemed
mistake was immediately to be expelled from the examiner profes- demonstrable, intuitive, and unencumbered by complicated statisti-
sion. The mistake thus would be removed from consideration as an cal and population genetic considerations.
error attributed to that community’s aggregate ‘product’. Cole Within a few years, however, forensic DNA ‘fingerprinting’ dis-
points out that such stringent discipline was not always, or even placed its former namesake, and was widely recognised even by
often, enacted in practice, but it served to protect the rule that former critics, such as Eric Lander, as the gold standard among
competent examiners never made errors: if they did err, this forensic tools.38 Ironically, the very characteristics that had been
deemed problematic in comparison with fingerprinting—the cum-

34
Goffman (1959), p. 2.
35
Cole (2005).
36
Lander (1989), Lewontin & Hartl (1991).
37
Aronson (2007).
38
See Lander & Budowle (1994).
 M. Lynch / Studies in History and Philosophy of Biological and Biomedical Sciences 44 (2013) 60–70 65

bersome statistical and population genetic estimates—were treated of the technique’.44 Though it too was presumptive in less obvious
as strengths. The extensive debates in admissibility hearings and ap- ways, the zero that was approached by extraordinarily low RMPs
peal cases about practical, analytical, and interpretative procedures was calculated in specific cases of forensic DNA analysis through
with DNA evidence now were viewed as evidence that ‘DNA’ had the use of officially approved methods. Although debates continued
passed the test, and could now be held reliable in terms of the about how to calculate RMPs, especially in the case of database ‘cold
new standards for the admissibility of expert ‘scientific’ evidence hits’, forensic analysts were able to invoke the formulae recom-
articulated by the U.S. Supreme Court in its 1993 ruling in Daubert mended in official audits of forensic DNA profiling by the U.S. Na-
v. Merrell Dow Pharmaceuticals, Inc.39 The statistical procedures and tional Research Council (NRC) and other organisations.45 The same
correction factors, which had seemed so obscure when compared audits recommended quality assurance/quality control standards
with the tidy declarations of fingerprint examiners, now stood as that could be cited as bases for putting to rest arguments about lab-
emblems of transparency, only now transparency did not mean intu- oratory error, even if it remained doubtful that such recommenda-
itively apparent; instead, it referred to calculations of probative value tions were rigorously followed. Judicial acceptance that such
that were logically traceable through formulae accepted by experts. administrative recommendations had been implemented was insep-
By comparison, the fingerprint examiner began to be stereotyped as arable from the technical fixes ‘themselves’.
a quaint figure: a ‘distinguished, greying individual on the stand say- Judicial decisions to admit DNA profile evidence in one jurisdic-
ing, ‘‘my opinion is based on my many years of experience in this tion after another in the U.S. effectively ended the widely publi-
field’’’.40 So, what accounts for this reversal of forensic fortune, or, cised disputes that had occurred in a relatively few cases in the
as we have elsewhere called it, this ‘inversion of credibility’?41 late 1980s and early ’90s. In addition, court decisions allowed
forensic analysts to cite RMP figures without combining them with
estimates of the probability of practical error, which were likely to
6. ‘Fixing’ a techno-legal controversy be much higher, if calculable at all.46 In the U.K., a number of legis-
lative initiatives and court decisions facilitated the collection of DNA
By the mid-1990s, criticisms of DNA profiling results were lar- profile evidence from suspects and non-suspects alike.47 Some of the
gely confined to charges of incompetence or bias levelled at partic- legislative initiatives were responsive to technical developments
ular police forces or forensic laboratories. The techniques that, for example, enabled profiles to be developed from small
‘themselves’ had become relatively immune to criticism. No single amounts of bodily sample (e.g., collected by cheek swabs rather than
technical solution, legal decision, or other event was responsible drawing blood) and DNA profiles to be digitised and stored on
for ending controversy in the courts and science press about foren- searchable databases. There was thus a developing reciprocal inter-
sic DNA profiling. Instead, a confluence of technical, legal, and action between technical and legal developments.
administrative ‘fixes’ worked to close debate about the reliability Administrative fixes were less visible than legislative or techni-
of the technique.42 cal changes, but more ubiquitous. Such changes often were initi-
There were several important ‘technical fixes’. The STR (short ated under the banner of ‘science’. As noted above, the NRC
tandem repeat) system replaced the earlier systems, and auto- reports made quality assurance/quality control (QA/QC) recom-
mated scanning of samples replaced visual examination of profile mendations for managing the collection, handling, and analysis
results. In theory, the STR system would register the exact sizes of evidence by police forces and forensic laboratories. Specific rec-
of the labelled alleles in a sample. That system also enabled ana- ommendations for using blank reagent controls and blind profi-
lysts to develop precise profile results from tiny amounts of bodily ciency testing, generating paper trails to track the movement of
material. As the system was established for widespread use in the samples, and using correction factors when calculating RMP fig-
U.K., U.S., and other nations, the number of DNA loci that were in- ures, produced standards of accountability that attorneys cited
cluded in a given profile increased from four to six, and then to and used when interrogating expert witnesses. Given the over-
eight and more. With each increase in the number of loci, the esti- whelming likelihood that criminal defendants would lack the re-
mated probability of coincidence matches decreased accordingly, sources to hire experts to observe the prosecution’s forensic
to the point that some forensic analysts argued that they should analysts at work or to perform their own analyses of the evidence,
be permitted in most cases to declare that the chance of a coinci- records and other paper assurances provided by the prosecution’s
dence match (for example, a match between a profile developed forensic analysts furnished virtual guarantees of adequate practice.
from a crime scene, and a DNA profile developed from a sample ta- Occasionally, when particular forensic laboratories came under
ken from a person unrelated to the ‘donor’ of the crime scene evi- intensive scrutiny, the picture that emerged was far from reassur-
dence) was, effectively, zero.43 Courts continued to require ing. A notorious example was the Houston Police Department
quantitative estimates of random match probability (RMP), but as Crime Laboratory, which was investigated in 2003 for an array of
the odds decreased from one in millions, billions, and even septil- problems with handling and interpreting forensic evidence.48 Pre-
lions, RMPs demonstrably approached zero. As a rhetorical matter, sumably, this was an exceptionally shoddy operation, but it re-
essentially zero RMP in cases of forensic DNA analysis differed from mained unclear what comparable scrutiny of practices in ‘normal’
the zero error rate claimed by proponents of latent fingerprint exam- forensic laboratories would reveal. Recommendations for handling
ination. The latter was presumptive, and it was preserved in the face such problems remained within the ambit of administrative fixes:
of errors that were invariably ascribed to human error and not ‘error more rigorous adherence to recommended protocols, more frequent

39
The U.S. Supreme Court majority in Daubert v. Merrell Dow Pharmaceuticals (1993) listed what they called ‘factors’ for trial courts to consider when deciding the admissibility
of (purported) scientific evidence: testability, peer review, measurable error rate and available standards, and general acceptance in the relevant fields.
40
Champod & Evett (2001), p. 106.
41
Lynch (2004), Lynch et al. (2008), p. 13.
42
Lynch et al. (2008), chapter 7.
43
Budowle, Chakraborty, Carmody, & Monson (2000).
44
Cole (2005).
45
NRC (1992, 1996); also see Cole & Lynch (2006).
46
Koehler (1996).
47
Lynch & McNally (2009).
48
Thompson (2003).
66 M. Lynch / Studies in History and Philosophy of Biological and Biomedical Sciences 44 (2013) 60–70

use of blind proficiency testing, higher standards for accreditation of also is to assert that the much-heralded objectivity of DNA profiling
forensic personnel, and so on. is inseparable from actual and presumed administrative
During early disputes about forensic DNA profiling, a technol- procedures.52
ogy transfer issue was prominent. DNA profile techniques were
used widely in numerous fields of scientific investigation, as well 7. Post-closure DNA
as in clinical diagnosis and paternity testing, prior to being applied
in criminal investigation. The question was whether they were, in For the past few decades, there have been debates in science
fact, the same techniques when implemented by police employees studies about how or why scientific controversies end or ‘close’.53
and contracting agencies and companies. It was widely assumed The classic realist position is that controversies end when the rele-
that ‘the same’ techniques were highly reliable when deployed in vant groups of scientists test the competing theories and deter-
‘science’. However, as a forensic analyst pointed out in an inter- mine—either through a crucial test or an accumulation of
view, the use of some of the constituent techniques may turn out mutually-supportive observational and experimental results—that
to be less reliable in contexts of innovative research, which often the evidence supports one theory but not the other (or others). Social
involve less routinised procedures, less administrative oversight, constructionists argued that evidence alone does not (and, indeed,
and less dramatic consequences for error than in clinical and foren- cannot) determine closure, and that a variety of ‘social’ conventions
sic circumstances.49 and commitments enter into the alignments of a ‘core group’ of sci-
Some of the admissibility hearings in the U.S. in the late 1980s entists around particular facts and theories.54 Although it is widely
and early ’90s addressed this transfer problem in terms of the assumed that the constructionist position is consonant with a socio-
question of who could speak for the reliability of DNA profiling. logical analysis, it is especially obvious in the case of a techno-legal
Proponents of forensic DNA analysis sometimes argued that, given controversy that conceptions of controversy, closure, and social con-
the specificity of forensic DNA profiling, the most authoritative struction are second order phenomena.55 By this is meant that these
witnesses for ‘representing’ the technique should be experienced conceptions are embedded in the disputes they describe, and ques-
employees of public and private labs who actually worked with tions on which fields and which scientists count as the ‘core group’
criminal evidence. Critics tended to argue that the testimony of of scientists are themselves adjudicated by the participants in the
‘scientists’ (molecular biologists, population geneticists, statisti- disputes, especially judges.56
cians) who had no direct stake in forensic science should take pre- The techno-legal controversy over DNA profiling played itself
cedence over the (usually less credentialed) employees of private out through the courts, with important but not exclusive input
and public forensic laboratories. A key judicial ruling in New York from judicially recognised ‘experts’ from several different fields.
v. Wesley (1988) established that expert witnesses should include Parties to the controversy themselves debated about what counted
a broader array of specialities in addition to forensic science.50 Con- as controversy and closure. Claims that there was controversy
sequently, arguments among the ‘scientists’ who took part in these about DNA profiling were often denied by proponents of the tech-
techno-legal debates focused on the reliability of forensic DNA profil- nology, who argued that there was no real controversy about DNA
ing, and panels such as the two NRC committees included many profiling but only disputes in specific cases about technical errors
prominent scientists, lab administrators, and legal scholars who and interpretations. Announcements that the controversy was
had no direct experience with forensic work. closed (or ‘put to rest’ in the words of Lander and Budowle) were
Although judges took advice from expert authorities who had themselves contested.57 Perhaps it would be fair to say that the
published technical evaluations of the constituent procedures, ‘controversy’ was simultaneously a ‘meta-controversy’: a dispute
courtrooms provided the most consequential venue for resolving about whether the ongoing dispute was a genuine controversy or a
debates about the reliability and validity of DNA typing. In key ‘controversy’ cooked-up by defence lawyers and civil libertarians.
courtroom ‘tests’ (and even in the NRC reports), the adequacy of By and large, parties who took a realist view of DNA profile results
‘scientific’ uses of DNA profiling was largely presumed, while foren- took a sceptical view of continuing controversy about such results,
sic uses of the relevant techniques were called into question. and parties who expressed scepticism about such results tended to
Administrative QA/QC recommendations apparently bridged the accept that the controversy was genuine. However, by the late
gap between scientific adequacy and questionable forensic prac- 1990s, even Richard Lewontin, who remained critical of how legal
tice, even though it remained unclear if, and to what extent, blind and administrative closure was reached, acknowledged that the con-
proficiency testing, blank reagent controls, and other such proto- troversy had ended for all practical purposes.58 The game was over,
cols were ‘rigorously’ adhered to in ‘basic science’ labs. The admin- even if there were lingering complaints about the referee’s decisions.
istrative recommendations for forensic labs in the U.S. and U.K.— Often, stories about controversies end when the controversies
such as precautions against contamination and mislabelling of close. In the case of forensic DNA profiling, however, the ‘post-clo-
samples, and the use of bar codes and scanners to track samples sure’ phase was no less interesting than the more contentious
through chains of custody—were more akin to those deployed in phases.59 In many public pronouncements, the various methods of
large volume diagnostic labs than in university laboratories.51 This DNA profiling were glossed with the simple acronym ‘DNA’. ‘DNA’
is not to discount the importance of such administrative procedures, became an evidential superhero: a powerful, even unassailable,
but to mark their specificity in a particular techno-legal context. It agent of truth and justice. Aside from being the subject of the

49
Jordan & Lynch (1992, 1993).
50
Aronson (2007), pp. 42–54.
51
Jordan & Lynch (1993).
52
See Cambrosio, Keating, Schlich, & Weisz (2006) for a discussion of ‘regulatory objectivity’: an ideal expressed in clinical trial designs that contend with a complex of
experimental, ethical, and pragmatic contingencies.
53
Collins (1985), Galison (1987).
54
Collins (1985).
55
Winch (1958), Schutz (1964).
56
Edmond (2001).
57
Lander & Budowle (1994).
58
Richard Lewontin, interviewed by Kathleen Jordan (7 April 1998).
59
Lynch et al. (2008), chapter 8.
 M. Lynch / Studies in History and Philosophy of Biological and Biomedical Sciences 44 (2013) 60–70 67

above-quoted quasi-religious pronouncements about ‘God’s signa- tion’s evidence included the victim’s eyewitness testimony as
ture’ and the ‘revelation machine’, DNA profiling provided an episte- well as DNA profile results that later were criticised for being mis-
mic lever with which to move some weighty legal institutions. interpreted. The case was reopened following investigative report-
Statutes of limitations were one such institution. Many U.S. ing by a local television station, and pursued by an official
states mandate that prosecutions for particular crimes must be ini- investigation of the case along with many other cases processed
tiated within a specified number of years after the crime occurs. by the Houston County forensic lab.65 Reanalysis of the bodily evi-
The rationale for such statutes of limitation for some categories dence that had been used to convict Sutton was used to exonerate
of crime (for example, some sexual offences) is that eyewitness’ him. After his release, the victim continued to maintain that he was
recollections are likely to degrade over time, and the passage of one of her assailants. According to a news report, the prosecutor con-
years makes it less and less likely that victims would correctly rec- tinued to back his witness in the face of the DNA reanalysis:
ognise their assailants. Similarly, bodily evidence would be likely
‘If this is not categorically dispositive evidence of innocence,’
to degrade over time, to the point of being useless for evidentiary
said David Dow, a lawyer for Mr. Sutton, referring to the results
purposes. However, ‘DNA’ has been granted exceptional treatment
of the DNA testing of sperm collected from the victim, ‘there is
in this context. In some jurisdictions, statutes of limitation were re-
no such thing.’ Chuck Rosenthal, the district attorney in Hous-
pealed in cases in which DNA evidence was available. In other
ton, saw things differently. ‘From the standpoint of the law,
jurisdictions, prosecutors devised a method for indicting a DNA
he’s innocent until proven guilty,’ he said. ‘Whether he’s actu-
profile in order to delay the lapse of statutes of limitation in cases
ally innocent, I don’t know,’ he said. ‘I’m not about to call the
in which no person had yet been indicted.
victim in this case a liar.’66
New York State law requires that a felony prosecution be
According to the same news report, prosecutors in the state of
brought within five years of those crimes, or within 10 years
Florida were supporting a time limit for re-opening closed cases,
if the criminal’s identity is unknown. The purpose of the limita-
and some were claiming that the use of DNA evidence to ‘prove’
tion is to protect the accused against fading memories and lost
innocence was in many cases ‘overstated’. For example, in court
witnesses. An indictment, even one that identifies the attacker
papers, a prosecutor in a Florida case ‘emphasized the justice sys-
simply by DNA—usually in semen collected soon after a
tem’s interest in finality, the hardship that a retrial would inflict on
crime—indefinitely preserves the ability to prosecute, officials
the victim and the strength of the remaining evidence. ‘‘The fact
said.60
that it can now be said that the defendant was not the source of
Two aspects of currently used DNA profiling methods were sig- the hair has little significance,’’ he wrote’. Critics of such pro-
nificant in such cases. One was that the STR system was deemed nouncements noted that prosecutors often were more sanguine
capable of developing usable results from old—and even an- about DNA testing when (as in many cases) it supported their argu-
cient—bodily remains. The other was that, except in cases of clo- ments for conviction. But, when resisting re-testing, prosecutors
sely related persons, RMP estimates were deemed sufficiently voiced arguments that defence attorneys often were faced with
close to zero that an indictment of a profile would stand proxy making in criminal trials about how the probative value of DNA
for an indictment of a unique individual ‘named’ by a series of tests was contingent upon the circumstances of the crime and
markers (one indictment in California listed a series of profile investigation:
markers and an RMP of ‘1 in 21 sextillion of the Caucasian
. . . prosecutors concede that DNA testing of the sperm found in a
population’.61
rape of an 11-year-old girl shows that it could not have come
The exceptional status granted to DNA evidence implicated
from the defendant [Mr. McKinley]. The prosecutors now say
other forms of evidence. Analysts of forensic science who had once
that doesn’t matter. ‘The DNA is a sideshow,’ said Edward Grif-
been critical of DNA profiling, began to use ‘DNA’ as a baseline
fith, a spokesman for the Miami-Dade state attorney’s office.
measure of the fallibility of other forms of evidence.62 In addition,
More important than the DNA, Mr. Griffith said, was a police
courts began to take a more favourable view of psychologists’ testi-
officer’s testimony that he saw Mr. McKinley atop the girl with
mony about the fallibility of eyewitness testimony.63 Cases in which
his pants down. The DNA evidence does not contradict that tes-
DNA evidence contradicted eyewitness testimony—even testimony
timony, he added, because the girl had had sex with another
expressed with great personal conviction—underlined what experi-
man not long before the rape. She also said that Mr. McKinley
mental psychologists such as Elizabeth Loftus had been saying for
had not ejaculated. Testimony about the girl’s earlier sexual
decades: that despite its privileged place in the legal system, eyewit-
encounter was excluded under a Florida law that bars introduc-
ness testimony was subject to suggestion, faulty recollection, and
tion of a victim’s sexual history in a rape trial. ‘That’s insane,’
other sources of fallibility.64
said Mr. Scheck, who represents Mr. McKinley in his current
Although prosecutors often are strong proponents of DNA profil-
request. ‘Whoever had sex with an 11-year-old committed a
ing—given the fact that it is commonly used in the investigation and
crime.’ He said that Mr. McKinley deserved at least a retrial.
prosecution of crimes—they sometimes resist its use for re-opening
‘There is no question that the prosecution took the position at
‘closed’ cases. Efforts to reopen such cases have provided dramatic
trial that the semen came from McKinley,’ he said.67
confrontations between ‘DNA’ and other forms of evidence. One
case in point was the Josiah Sutton case in Houston. Sutton, a young In an interview, perhaps alluding to this case as well as oth-
African American man, had been convicted for a 1998 rape involv- ers, Barry Scheck and Peter Neufeld spoke of the ‘unindicted
ing two assailants and sentenced to a 25-year term. The prosecu-

60
Rashbaum (2003).
61
Starrs (2000), p. 4.
62
Saks & Koehler (2005).
63
Dorf (2001).
64
Loftus (1975). See Winter (2013) for historical analysis.
65
Thompson (2003).
66
Liptak (2003).
67
Ibid.
68 M. Lynch / Studies in History and Philosophy of Biological and Biomedical Sciences 44 (2013) 60–70

co-ejaculator’ argument (recalling the memorable phrase ‘unin- would degrade if there were no end to the possibility of challenging
dicted co-conspirator’ from the U.S. Watergate hearings).68 legal outcomes. And, of course, endless challenges would present the
Similar complications arose in R v. Shirley (2003) the first appeal courts with an administrative nightmare. As the late Chief Justice
case in the U.K. in which post-conviction DNA evidence was used William Rehnquist argued, the threshold for reopening closed cases
to exonerate a man convicted of rape.69 The Crown prosecutor in ‘would necessarily be extraordinarily high’, because of ‘the very dis-
this case acknowledged that the DNA profile from the criminal evi- ruptive effect that entertaining claims of actual innocence would
dence did not match the defendant’s profile, but argued that the se- have on the need for finality in capital cases, and the enormous bur-
men evidence was from an unidentified man with whom the victim den that having to retry cases based on often stale evidence would
had engaged in consensual sex some hours earlier. This led to a fur- place on the States . . .’.75
ther dispute over the victim’s movements during the day and The promotion of DNA evidence as ‘a scientific arbiter of truth’
whether the perpetrator’s semen would have been recovered from that exposes the fallibility of all other forms of criminal evidence
the examination of the victim’s body and undergarments. Whether suggests an epistemic status that is an exception to the ‘human fal-
or not one agrees with such prosecutorial arguments, they point to libility’ that justifies administrative closure and moral certainty.76
the necessity to ‘contextualize the significance of any particular in- The doctrines that DNA does not degrade over time and that DNA
stance of DNA profiling’.70 However, the trend in recent years to ac- profiles are unique to individuals—the same doctrines that justify
cept ‘DNA’ as ground truth discourages such efforts to repealing or circumventing statutes of limitation when DNA evi-
contextualise.71 dence is available—also justify exceptions to the procedural and
The detailed arguments in these cases reveal that DNA testing temporal limits associated with finality.77 There is no question that
does not simply confirm or contradict other forms of evidence in finality can be invoked as a convenient excuse to avoid reopening
criminal trials, but that its probative value, like that of other evi- convictions gained on dubious grounds.78 The higher-than-moral
dence, is embedded in competing versions of the circumstances certainty associated with ‘science’ (with ‘DNA’ as its avatar) has pro-
of the crime. To use DNA test results as a baseline measure to as- vided strong leverage for the Innocence Project’s efforts to reopen
sess error rates of other forms of evidence not only presumes that closed cases, but a reciprocal possibility has not been lost on propo-
DNA is, for all practical purposes, error-free, but that the different nents of the death penalty in the USA. In 2003, Mitt Romney, who at
types of evidence can be categorically isolated from the circum- the time was Governor of Massachusetts, one of the relatively few
stances of crimes and investigations relayed through testimony U.S. states that had not legalised the death penalty, was quoted as
and documents. Although efforts to exonerate falsely accused saying, ‘[j]ust as science can be used to free the innocent, it can also
defendants are to be applauded, it is worth keeping in mind that be used to identify the guilty’.79 Romney envisioned what might be
much of the work involved in pursuing exoneration consists of re- called the ‘Guilty Project’: an expert panel that would review evi-
views of evidence and records of investigation that reveal prosecu- dence from certain heinous crimes, to decide on ‘scientific’ grounds
torial suppression of evidence, shoddy representation of indigent if the evidence of guilt was ‘incontrovertible’.80 Romney’s proposal
defendants, pressured testimony, and many other non-scientific is- did not receive legislative support, but it pointed to the possibility
sues. In brief, DNA evidence is testimony; it is not a gift from sci- that the exceptional degree of certainty ascribed to ‘scientific’ evi-
ence that remains apart from the contingencies of investigative dence in the criminal justice system could be used to justify the
and trial processes.72 selective use of the most final of penalties. If proposals like Romney’s
were to succeed, the same source of ‘scientific’ leverage against legal
8. Conclusion finality would be used to install a more permanent form of finality.81
Currently, DNA evidence is treated as exceptional, but it also is
The extraordinary truth-value attributed to ‘DNA’, following the upheld as a model for other forms of forensic evidence to emulate.
closure of controversy about it in the criminal courts and science Given the current state of forensic science, there is reason to wel-
press, has encouraged its use as leverage to reopen closed cases. come an expansion of precision estimation (expressed through
As shown above, prosecutors resisting post-conviction DNA testing probability figures), upgraded procedural standards and practi-
sometimes invoke the legal principle of ‘finality’. Finality refers to a tioner credentials, protections against error, and so forth. However,
procedural necessity to administer certainty in the face of an expli- there also is a reason to be concerned about the possible institu-
cit recognition that guilt and innocence are determined ‘within the tionalisation of a threshold of certainty ascribed to those technol-
limits of human fallibility’.73 ‘Finality’ is a formal, institutional ver- ogies that ‘pass the test’. Even if it makes sense to say that DNA
sion of the commonplace expression ‘enough is enough’.74 It is ex- evidence is ‘nature’s testimony’, it nevertheless remains the case
pressed through restrictions on a convicted defendant’s rights to that its probative value as testimony is bound up in competing
sue the defence attorney, or to reopen a case after grounds for appeal and all-too-human narratives: stories about how DNA and other
are deemed exhausted. It is justified on the grounds of moral cer- forms of evidence implicate ‘what happened’ at the scene of an al-
tainty—maintaining trust in the fairness and certainty of fallible legal leged crime. As such, it should not be exempt from moral certainty
judgments. The rationale is that public trust in legal institutions ‘within the limits of human fallibility’.

68
Barry Scheck and Peter Neufeld, interviewed by Simon Cole and Michael Lynch (21 May 2003).
69
Johnson & Williams (2004).
70
Ibid., p. 80.
71
Aronson & Cole (2009).
72
Koehler (1996).
73
Wainwright v. Sykes (1977).
74
Garfinkel (1967), p. 28.
75
Herrera v. Collins (1993).
76
Findley (2002), p. 336.
77
Berger (2004), p. 113.
78
Many such cases are described by Scheck et al. (2000).
79
Phillips (2003).
80
Swanson (2005), p. 12.
81
Aronson & Cole (2009).
 M. Lynch / Studies in History and Philosophy of Biological and Biomedical Sciences 44 (2013) 60–70 69

Acknowledgements Jordan, K., & Lynch, M. (1992). The sociology of a genetic engineering technique:
Ritual and rationality in the performance of the plasmid prep. In A. Clarke & J.
Fujimura (Eds.), The right tools for the job: At work in twentieth-century life science
Simon Cole, Ruth McNally and Kathleen Jordan participated in (pp. 77–114). Princeton, NJ: Princeton University Press.
the research projects that informed this paper. I am grateful for Koehler, J. J. (1996). On conveying the probative value of DNA evidence.
Frequencies, likelihood ratios, and error rates. University of Colorado Law
support from the ESRC (UK) and NSF (US) for those projects, partic-
Review, 67, 859–886.
ularly for the NSF award SES-0115305, ’DNA profiling and finger- Lander, E. S. (1989). DNA fingerprinting on trial. Nature, 339, 501–505.
printing: Relations between closure and controversy’. I also Lander, E. S., & Budowle, B. (1994). DNA fingerprinting dispute laid to rest. Nature,
371, 735–738.
would like to thank Ian Burney for his helpful comments and edi-
Latour, B., & Woolgar, S. (1979). Laboratory life: The social construction of scientific
torial advice on earlier drafts of this paper. facts. London: Sage.
Laudan, L. (2003). Is reasonable doubt reasonable? Legal Theory, 9, 295–331.
References Lewontin, R. C., & Hartl, D. J. (1991). Population genetics in forensic DNA typing.
Science, 254, 1745–1750.
Liptak, A. (2003). Prosecutors fight DNA use for exoneration. New York Times (29
Aronson, J. (2007). Genetic witness: Science, law, and controversy in the making of DNA August).
profiling. New Brunswick, NJ: Rutgers University Press. Locke, J. (1995[1690]). An essay concerning human understanding, Amherst, NY:
Aronson, J., & Cole, S. (2009). Science and the death penalty: DNA, innocence, and Prometheus Books.
the debate over capital punishment in the U.S.. Law & Social Inquiry, 34, Loftus, E. (1975). Leading questions and the eyewitness report. Cognitive Psychology,
603–633. 7, 560–572.
Barnes, B., Bloor, D., & Henry, J. (1996). Scientific knowledge: A sociological analysis. Lynch, M., Cole, S., McNally, R., & Jordan, K. S. (2008). Truth machine: The contentious
Chicago: University of Chicago Press. history of DNA fingerprinting. Chicago: University of Chicago Press.
Barnes, S. B., & Dolby, R. G. A. (1970). The scientific ethos: A deviant viewpoint. Lynch, M. (2004). ‘Science above all else’: The inversion of credibility between
Archives European Journal of Sociology, 11, 3–25. forensic DNA profiling and fingerprint evidence. In G. Edmond (Ed.), Expertise in
Berger, M. A. (2004). Lessons from DNA: Restriking the balance between finality and law and regulation (pp. 121–135). Aldershot: Ashgate.
justice. In D. Lazer (Ed.), DNA and the criminal justice system: The technology of Lynch, M., & McNally, R. (2009). Forensic DNA databases and biolegality: The co-
justice (pp. 109–131). Cambridge, MA: MIT Press. production of law, surveillance, technology and suspect bodies. In P. Atkinson,
Bloor, D. (1976). Knowledge and social imagery. London: Routledge. P. Glasner, & M. Lock (Eds.), Handbook of genetics and society: Mapping the new
Budowle, B., Chakraborty, R., Carmody, G., & Monson, K. (2000). Source attribution genomic era (pp. 283–301). London: Routledge.
of a forensic DNA profile. Forensic Science Communications, 2. March, J. G., & Simon, H. (1958). Organizations. New York: Wiley.
Cambrosio, A., Keating, P., Schlich, T., & Weisz, G. (2006). Regulatory objectivity and Martin, A. (2011). The prosecution’s case against DNA. New York Times (25
the generation and management of evidence in medicine. Social Science & November).
Medicine, 63, 189–199. Merton, R. K. (1970). Science, technology and society in seventeenth-century England.
Caudill, D., & LaRue, L. (2006). No magic wand: The idealization of science in law. New York: Harper & Row.
Lanham, MD: Rowman & Littlefield. Merton, R. K. (1973[1942]). The normative structure of science. In R. K. Merton (Ed.),
Caudill, D., & Redding, R. (2000). Junk philosophy of science? The paradox of The sociology of science (pp. 267–278). Chicago: University of Chicago Press.
expertise and interdisciplinarity in federal courts. Washington & Lee Law Review, Mitroff, I. (1974). Norms and counter-norms in a select group of the Apollo moon
57, 685–766. scientists. American Sociological Review, 39, 579–595.
Champod, C., & Evett, I. W. (2001). A probabilistic approach to fingerprint evidence. Mnookin, J. (2001). Fingerprint evidence in the age of DNA profiling. Brooklyn Law
Journal of Forensic Identification, 51, 101–122. Review, 67, 13–70.
Cole, S. A. (2001). Suspect identities: A history of fingerprinting and criminal Mulkay, M. (1976). Norms and ideology of science. Social Science Information, 15,
identification. Cambridge, MA: Harvard University Press. 637–656.
Cole, S. A. (2005). More than zero: Accounting for error in latent fingerprint NRC (National Research Council). (1992). DNA technology in forensic science.
identification. Journal of Criminal Law and Criminology, 95, 985–1078. Washington DC: National Academy Press.
Cole, S. A. (2013). Forensic culture as epistemic culture: The sociology of forensic NRC (National Research Council). (1996). The evaluation of forensic DNA evidence.
science. Studies in the History and Philosophy of Biology and Biomedical Sciences, Washington DC: National Academy Press.
44, 36–46. Phillips, F. (2003). Science stressed on death penalty. Boston Globe, p. 1 (24
Cole, S. A., & Lynch, M. (2006). The social and legal construction of suspects. Annual September).
Review of Law and Social Science, 2, 39–60. Popper, K. (1959). The logic of scientific discovery. London: Hutchinson & Co.
Collins, H. M. (1985). Changing order: Replication and induction in scientific practice. Popper, K. (1963). Conjectures and refutations. London: Routledge & Kegan Paul.
London: Sage. Prainsack, B., & Kitzberger, M. (2009). DNA behind bars: Other ways of knowing
Dorf, M. C. (2001). How reliable is eyewitness testimony? A decision by New York forensic DNA technologies. Social Studies of Science, 39, 51–79.
State’s highest court reveals unsettling truths about juries. Findlaw, legal news Prainsack, B., & Toom, V. (2013). Performing the union: The Prüm decision and the
and commentary. Available at <http://writ.news.findlaw.com/dorf/ European dream. Studies in the History and Philosophy of Biology and Biomedical
20010516.html>. Sciences, 44, 71–79.
Edmond, G. (2001). The law-set: The legal-scientific production of medical Rashbaum, W.F. (2003). New York pursues old cases of rape based just on DNA. New
propriety. Science, Technology, & Human Values, 26, 191–226. York Times (5 August).
Edmond, G. (2002). Legal engineering: Contested representations of law, science Rimer, S. (2002). Convict’s DNA sways labs, not a determined prosecutor. New York
(and non-science), and society. Social Studies of Science, 32, 371–412. Times (6 February).
Edmond, G., & Mercer, D. (1997). Scientific literacy and the jury: Reconsidering jury Saks, M. J., & Koehler, J. J. (2005). The coming paradigm shift in forensic
‘competence’. Public Understanding of Science, 6, 329–357. identification science. Science, 309, 892–895.
Findley, K. (2002). Learning from our mistakes: A criminal justice commission to Scheck, B., Neufeld, P., & Dwyer, J. (2000). Actual innocence. Five days to execution,
study wrongful convictions. California Western Law Review, 38, 333–354. and other dispatches from the wrongly convicted. New York: Doubleday.
Galison, P. (1987). How experiments end. Chicago: University of Chicago Press. Schutz, A. (1964). The problem of rationality in the social world. In Collected papers
Garfinkel, H. (1967). Studies in ethnomethodology. Englewood Cliffs, NJ: Prentice II: Studies in social theory (pp. 64–90). The Hague: Martinus Nijhoff.
Hall. Shapin, S. (1994). The social history of truth. Chicago: University of Chicago Press.
Garfinkel, H., & Sacks, H. (1970). On formal structures of practical actions. In J. C. Shapin, S. (1995). Cordelia’s love: Credibility and the social studies of science.
McKinney & E. A. Tiryakian (Eds.), Theoretical sociology: Perspectives and Perspectives on Science, 3, 255–275.
development (pp. 337–366). New York: Appleton-Century-Crofts. Shapin, S., & Schaffer, S. (1985). Leviathan and the air pump. Princeton, NJ: Princeton
Gieryn, T. (1983). Boundary-work and the demarcation of science from nonscience. University Press.
American Sociological Review, 48, 781–795. Shapiro, B. J. (1986). ‘To a moral certainty’: Theories of knowledge and Anglo-
Gieryn, T. (1999). Cultural cartography: Credibility on the line. Chicago: University of American juries, 1600–1850. Hastings Law Journal, 38, 153–193.
Chicago Press. Shapiro, B. J. (1991). ‘Beyond reasonable doubt’ and ‘probable cause’: Historical
Goffman, E. (1959). The presentation of self in everyday life. Garden City, NY: perspectives on the Anglo-American law of evidence. Berkeley, CA: University of
Doubleday Anchor. California Press.
Haack, S. (2005). Trial and error: The supreme court’s philosophy of science. Shapiro, B. J. (2002). Testimony in seventeenth-century English philosophy: Legal
American Journal of Public Health, 95(S1), S66–S73. origins and early development. Studies in History and Philosophy of Science, 33,
Jasanoff, S. (1995). Science at the bar: Law, science, and technology in America. 243–263.
Cambridge, MA: Harvard University Press. Solomon, S., & Hackett, E. (1996). Setting boundaries between science and law:
Johnson, P., & Williams, R. (2004). Post-conviction DNA testing: The U.K’.s first Lessons from Daubert v. Merrell Dow Pharmaceuticals, Inc.. Science, Technology, &
‘exoneration’ case. Science and Justice, 44, 77–82. Human Values, 21, 131–156.
Jordan, K., & Lynch, M. (1993). The mainstreaming of a molecular biological tool: A Starrs, J. E. (2000). The John Doe DNA profile warrant. Scientific Sleuthing Review, 24, 4.
case study of a new technique. In G. Button (Ed.), Technology in working order: Swanson, K. (2005). Certain guilt: The social construction of a scientific death penalty
Studies in work, interaction, and technology (pp. 160–180). London and New statute. Unpublished paper. Department of History of Science, Harvard
York: Routledge. University.
70 M. Lynch / Studies in History and Philosophy of Biological and Biomedical Sciences 44 (2013) 60–70

Thompson, W.C. (2003). Review of DNA evidence in State of Texas v. Josiah Sutton Cases
(District Court of Harris County, Cause No. 800450). Available at <http://
scientific.org/DNAProblems/DNA-Problems.htm#Houston%20background>.
Daubert v. Merrell Dow Pharmaceuticals, Inc., 509 U.S. 579 (U.S. 1993).
Weber, M. (1968). Economy and society: An outline of interpretive sociology. 2 Vols.
Herrera v. Collins, 506 U.S. 390, 408 (U.S. 1993).
Berkeley, CA: University of California Press.
Kitzmiller v. Dover Area School District, et al., 400 F.Supp. 2d 707 (M.D. PA 2005).
Weber, M. (1976). The protestant ethic and the spirit of capitalism (T. Parsons, Trans.).
McLean v. Arkansas Board of Education, 529 F. Supp. 1255 (E.D. Ark. 1982).
New York: Charles Scribner & Sons.
People v. Jennings, 96 N.E. 1077 (Ill. 1911).
Wilgoren, J. (2002). Confession had his signature: DNA did not. New York Times, p. A-
People v. Sallow, 100 Misc. 447, 165 N.Y.S. 915 (N.Y.C. Gen. Sess. N.Y. County 1917).
13 (26 August).
People v. Wesley, 140 Misc. 2d 306 (Albany Cty. Ct., NY, 1988).
Winch, P. (1958). The idea of a social science and its relation to philosophy. London:
Regina v. Shirley, Court of Appeal (Criminal Division), London, 29 July 2003.
Routledge & Kegan Paul.
Wainwright v. Sykes, 433 U.S. 72, 90 (U.S. 1977).
Winter, A. (2013). The rise and fall of forensic hypnosis. Studies in the History and
Philosophy of Biology and Biomedical Sciences, 44, 26–35.
Zimmerman, D. H., & Pollner, M. (1970). The everyday world as a phenomenon. In J.
D. Douglas (Ed.), Understanding everyday life: Toward the reconstruction of
sociological knowledge (pp. 80–103). Chicago: Aldine.