Anyone who has tried to make sense of the writings of Jacques Lacan may have puzzled over the phrase “welcome to the Desert of the Real,” intoned by the character Morpheus in the 1999 film The Matrix. The orgiastic scene at the beginning of the 2003 sequel, Matrix Reloaded, suggests that the desert is not to be found in real life but in the symbolic order of the matrix—in the code.

Richard Biernacki’s Reinventing Evidence in Social Inquiry posits cultural computer coding as a desert of the real. Cultural sociologists’ coding is for Biernacki a kind of textual torture. In the act of coding discourse, a text’s organically interwoven signifiers, sentences, and paragraphs are torn out of their original contexts, destroying any hope of understanding the text’s meaning. The resultant product consists of invented facts, false confessions. The mass of encoded texts threatens to become a “toxic” monster, a “thing-like structure” (18, 20)^{Footnote 1}.

Reinventing Evidence begins by calling attention to emerging methods in cultural sociology that claim to have overcome the age-old division between interpretivism and science through the formal coding of large numbers of texts. “When applied to qualitative data,” such quantitative methods claim to permit generalizations, thereby avoiding the putative “relativism and subjectivism of earlier interpretive sociologies.” As Biernacki observes, “whoever refers to their work as ‘coding’ rather than as mere reading is making a claim to participate in a scientific enterprise” (6-7). Biernacki’s aim is to be “divisive” in the “positive sense of splitting two .. universes of intellectual practice apart, that of quantitatively inspired sampling and variable analysis […] versus philological close reading and synthesis in the spirit of the humanities” (9). By then attempting to replicate the results from three well-known sociological studies that rely on cultural coding, Biernacki tries to demonstrate that the “humanities” approach is actually more “rigorous” (3). In calling the bluff of the scientistic camp Biernacki is brandishing one of the most powerful weapons in the arsenal of the humanistic critique.

The book’s core theoretical argument is that culture and meaning are defined contextually, that is, within specific texts or groups of texts. The cultural analyst is therefore compelled to leave intact the internal “textual interconnections” and “system-like relations that let us generate meaning” (137), rather than ripping symbols “out of their original contexts” and bringing them “into previously unimaginable relations to each other” (11) via “mathematical correlations” (9). Texts are able to signify not just in a positivist sense but also “by implicit contrast to what is not mentioned as well as by superseding or negating what they first affirm” (97). Connotation constitutes a second-order layer of signification, as Barthes argued.^{Footnote 2} The Russian formalists showed how narrative disjunctures between “story” (fabula) and “plot” (syuzhet) generate meaning.^{Footnote 3} Freud argued that dream symbols may only be comprehensible as condensations or displacements of other signifiers. Even the simplest signs may not make sense when severed from their textual social or historical context. Coders who assume that meaning is fixed—that the signifier “cigar” is always a cigar—are committing a very basic semiological error. Biernacki’s critics miss this straightforward point, believing him to be arguing that “words and phrases have no meaning outside their textual context.”^{Footnote 4}

Obviously a word or phrase divorced from its original context can have meanings of some sort. Biernacki’s point is that sociologists who engage in coding claim to be shedding more light on the meaning of the original text but are in fact putting those signs into entirely new contexts determined by the researcher and changing their meaning. The initial decontextualization via coding is followed by a recontextualization, as “meaningful texts” are turned “into unit facts for the sake of converting those units back into meanings.” Codes tend to find what coders are looking for, in a “self-confirming ritual” (8) that relies on “circularly self-fulfilling procedures for generating ‘findings’” (5). The “distinction between pattern discovery and pattern creation” is routinely suppressed (12). Indeed, some coders embrace a kind of perspectival idealism, arguing that “there can be no such thing as intercoder reliability, because each individual receives a unique observational world”, or comparing the researcher’s “project[ing] an interpretation” onto the coded data to “a Rorschach test” (7, 14).

The alternative is the humanities’ “more disciplined if less technical inquiry into cultural meanings” (17): meticulous, painstaking archival history and careful close readings. Biernacki focuses on seemingly simple practices like footnoting, sharing documentation, and subjecting ones own interpretations to “an acute trial” (3). In this approach, evidence “has the potential to challenge our preconceptions of what texts signify and of how they were put together” (16). Central to Biernacki’s version of this “countertradition” (16) is the Weberian ideal type, which is always open to “overhaul by anomalous evidence” and which allows the researcher “to highlight and retest personal case judgments about what is comparable [rather] than to keep these choices between the scenes as arbitrary unknowns” (5, 17). Biernacki quotes Musil to the effect that “every case on which thoughtful investigators land has the ability to overturn everything that people had up to then believed” (155), and singles out Arendt’s analysis of Eichmann as upending “everything we had previously supposed about a culture or ourselves” (136). Biernacki describes the practitioners of this countertradition as craft workers (3)—presumably as opposed to the industrial style of social science.^{Footnote 5}

Biernacki’s arguments in the first and final chapters bookend three case studies in which he tries to replicate three canonical works of cultural sociology: Bearman and Stovel’s “Becoming a Nazi,” Evans’ Playing God, and Griswold’s “The Fabrication of Meaning: Literary Interpretations in the United States, Great Britain, and the West Indies.” What he shows in these replications is just how decisively the protocols of cultural coding can shape the results. Bearman and Stovel code narratives of Nazis gathered at the urging of the Polish-American sociologist Theodor Abel in 1934. They map these narratives into network diagrams made up of nodes and connections. Bearman and Stovel report that their “analysis of the structure of NSDAP narrative networks suggests that the new Nazi self emerges from the elision of social relations.” Their article focuses on the case of Herr D., to whom they attribute the syndrome of a “master identity formation” that “gives rise to action which is insensitive to context.”^{Footnote 6} The emergence of this identity type is characterized by “decreasing centrality of relational elements and traditional bases of identity (e.g., kin, church, school, and work) in the narrative of becoming.”^{Footnote 7}

Biernacki retrieved the original Nazi life stories from the Hoover Institution archives and finds a serious of jarring misinterpretations. First, Bearman and Stovel’s coding relied on an abridged translation by Abel of the original German narrative. Using such a selective condensation of the narrative violates “the researchers’ justification for their investigation,” which was that the full autobiographies “represent the original writers’ personal arranging of story elements for coherence” (30). Given his own theoretical interests, Abel “was prone to deleting the musings on personal life decisions or self-reflection on family relations and social relations”—“the very stuff that Bearman and Stovel report as strangely missing and as indicating an absence of selfhood when Herr D. acquires a Nazi identity” (30). Biernacki shows that Herr D.’s original, full narrative illustrates that he had “intense family ties” during his time as a Nazi (47) rather than “drifting in a world stripped of social relations.”^{Footnote 8} Biernacki points out that Bearman and Stovel disregarded “massive evidence from historians about the multiplex social ties of Nazis before 1933,” who were “not drifters but opinion leaders entwined in community organizations” (35). Bearman and Stovel suggest that anti-Semitism “is characterized by low centrality,”^{Footnote 9} but as Biernacki observes, “Jews are pivotal in the autobiography, because precocious revulsion against them binds Herr D.’s childhood to his adult wisdom.” Indeed, he reasons, “vehemently removing Jews from your life would likely lead to their network marginality, all the better to function as an implicit focal contrast to Nazi values” (42). The coding, Biernacki concludes, “reverses the basic meaning of the story” (46). Indeed, techniques like this cannot capture narrative at all, which is much more than a sequence of events.”^{Footnote 10}

In one respect Biernacki’s dismantling of this article doesn’t really require the critique of coding, since these mistakes would be equally serious if they were found in a “humanistic” account. His reanalysis of Griswold’s “Fabrication of Meaning” focuses on coding per se. This study compared the different meanings that literary reviewers “from the West Indies, Britain, and the United States constructed from a single source, the fiction of Barbadian novelist George Lamming.”^{Footnote 11} After painstakingly reconstructing the lost sample of 95 reviews, Biernacki focuses first on Griswold’s argument that the novels that generated the most favorable reviews were those understood as being moderately ambiguous. Biernacki finds, however, that ambiguity is specifically mentioned “primarily when the reviewer expresses frustration or disappointment” with a text (1010). He concludes that the “entire test of the hypothesis depends on oracular coding results assumed to be easy and given, although they are beset by dilemmas” (106). Griswold also counted topics such as “race,” “colonialism,” “class,” and “revolution.” Again, these codes seem to collapse, since Caribbean reviewers “mocked the British predisposition to reify race as a thing, whereas for those who lived on the islands, race seemed […] a dimension inextricable from many others in the experience of colonial subjects” (112). As for colonialism, Griswold suggests that British critics “indicated their preoccupation with colonialism by avoiding the subject so persistently and by concentrating on style rather than content.”^{Footnote 12} Biernacki responds that if the same sociologist “can treat frequent mentions as evidence for a topic’s weight, as with “race” […] but reverse gears to treat infrequent mentions as evidence for a topic’s weight, as with “colonialism” […] then guessing the meaning of the numbers” is little more than a “Rorschach test” (114). In addition to these issues of coding, Biernacki questions whether the reviews are really commensurable units, bringing up a topic that receives fuller treatment in his analysis of the third study, Playing God.

Evans’ book analyzes articles and books on human genetic engineering (HGE) between 1959 and 1995 to show that they shifted from Weberian value rationality to formal rationality and to argue that this change was driven by the preferences of increasingly powerful US federal advisory committees.^{Footnote 13} Biernacki argues that texts often blend the two forms of rationality. But his main criticism concerns Evans’ sample construction procedures. Biernacki criticizes Evans’ redefinition of his criteria after 1973 away from philosophical references toward more specific, technologically sensitive keywords, noting that this sampling procedure excludes substantively rational works, in what amounts to self-fulfilling “narrowing of content over time” (63). Even a humanistic sociologist would need to decide which texts to include and which to exclude. Biernacki argues that the selection process should at least be made visible. By suggesting that some cultural objects are artificially created via sampling and simply don’t exist in the real world (and that HGE may be one such topic; 67), Biernacki hints at a more realistic definition of fields of cultural production as an alternative to random or subjectively arbitrary approaches to sampling. Biernacki also calls for various sorts of comparison: between HGE and other moral debates around topics like abortion; between the US and other national or regional contexts; and with arenas in which there are no federal advisory committees, so that one can assess their alleged impact on ethical discourse (61).

I cannot confirm or reject most of Biernacki’s counter-interpretations of the evidence without visiting the archives myself, but anyone familiar with Biernacki’s other work, especially his masterly The Fabrication of Labor,^{Footnote 14} knows that he is a fastidious historical and cultural researcher. One might object to Biernacki’s small sample of just three studies, but he deflects such criticism with his depiction of sampling in the humanities as an “artful rendering” that lays itself open for inspection and correction. It is also worth noting that Biernacki does refer to other examples of cultural coders, without engaging in full-scale reanalysis.

I am less convinced by Biernacki’s framing of social science as ritual. Biernacki argues that “only in ritual does the universal message evidenced in an encoding procedure replace the evidence itself as the sacred” (109). Each of the three case studies is supposedly organized around a storyline moving from (1) “the ritual phase of separation” of culture from its context, to (2) a “liminal phase” in which textual fragments are placed on the operating table (a “ritual platform”) in order for “simulacral facts” (117) to be extracted, and concluding with (3) a “phase of integration.” Biernacki’s interpretation of the final “phase of integration” relies on the idea that sociologists hope to restore a sacred, “unified world of obviousness and consensus by invoking ‘society’ as a secure […] referent” explaining culture (123). Equally plausible frameworks might frame coding as a labor process or cultural sociologists as embroiled in competitive scuffles within disciplinary subfields. Put differently, why would anyone look to sociology (rather than to high culture, for example) for a re-enchantment of the world? Is ritual really a general frame for all forms of social practice? If so, how does Biernacki’s own archival and text-exegetical work avoid being configured as sacred ritual?

A related problem is that Biernacki’s own epistemology and metaphysics remain murky. He suggests that interpretation and explanation are opposing and incompatible goals (7), but he notes that the Weberian ideal types he endorses do “the work of explaining” (17). He combines calls for comparison with insistence on the singularity of cultural facts, but he avoids Weber’s (and Rickert’s) language of “historical individuals.” He refers to “real entities” (17), causal “mechanisms” (109) and an Aristotelian realism of causal powers (2) without ever elaborating his own views on causality. Biernacki provides support for the critique of Humeanism in the social sciences by showing the proximity between empirical positivism and relativistic idealism, but without signaling whether that is his intention.^{Footnote 15} Nor does he clarify his position in the debate between Saussurians and Derridians. And Biernacki could have extended his critique to people like the sexologists of jouissance who attempt to code the non-symbolic or the non-discursive.^{Footnote 16}

To come finally to the epigraph that begins this review, it is only fitting that Brecht wrote the American version of his Galileo in Southern California, for that is where Richard Biernacki was subjected to a display of academia’s contemporary “instruments.” After complaints of being “harassed” made by another Professor (indicated in the report as Professor B) his Dean required him to “stop contacting B with questions regarding [name of B's publication], his/her research methods, or his/her previous research methods; stop contacting others about your re-analysis of his/her data; refrain from discussing [...] your re-analysis of B’s data at your presentations at any meetings, including scholarly meetings like the [name of professional association]; and do not publish texts that refer to [...] your re-analysis of B’s data.” If he did not comply, he was threatened with “written censure, reduction in salary, demotion, suspension or dismissal.” After a faculty investigation, the Dean’s letter was rescinded and the threats withdrawn.^{Footnote 17} Biernacki thus begins his book by slyly noting the similarities between the coders in social science and the “elevated point of view like that of an administrator” (4). The Biernacki affair should serve as a cautionary tale to those who believe that academic fields are not also battlefields. Reinventing Evidence is an immensely serious book that should be read by anyone involved in the business of sociology.