Introduction: Leaving the
Two Cultures Behind
A Question of Methods
Scientists typically
operate by formulating testable hypotheses and producing data to test the
hypotheses. Students of literature, in contrast, typically proceed by way of
argument and rhetoric. In their most scholarly guise, they aim at producing
objective textual and historical information, but all such information must
ultimately be interpreted within some larger order of ideas. During roughly the
first two thirds of the twentieth century, the most prominent theoretical
systems with a scientific cast were Marxist social theory, Freudian and Jungian
psychology, and structuralist linguistics. Even in their own fields, these
systems were only quasi-scientific, more speculative than empirical, and in
literary study, speculative ideas served chiefly as sources for imaginative
stimulation. Most critics operated as eclectic free agents, spontaneously
gleaning materials from every region of knowledge—from philosophy, the
sciences, history, the arts, and especially from literature itself. Though
using selected bits of information from the sciences, students of literature
commonly regarded their own kind of knowledge—imaginative, subjective,
qualitative—as an autonomous order of discourse incommensurate with the
quantitative reductions of science. Over the past three decades or so, these
older forms of literary criticism have been superseded by a new theoretical
paradigm designated variously as poststructuralism or postmodernism. The new
paradigm incorporates psychoanalysis and Marxism in their Lacanian and
Althusserian forms, but poststructuralists explicitly reject the idea that
scientific methods secure the highest standard of epistemic validity. Instead,
they include science itself within the rhetorical domain formerly set aside as
the province of the humanities.
As literary culture has
been moving away from the canons of empirical inquiry, the sciences have been
approaching ever closer to a commanding and detailed knowledge of the subjects
most germane to literary culture: human motives, human feelings, and the
operations of the human mind. Evolutionary psychology and affective
neuroscience have been penetrating the inner sanctum of the “qualitative” and
making it accessible to precise empirical knowledge. Since the early 1990s, some few literary scholars have
been assimilating the insights of the evolutionary human sciences and
envisioning radical changes in the conceptual foundations of literary study.
These “literary Darwinists” have produced numerous theoretical and interpretive
essays. Until recently, though, most literary Darwinists have remained within
the methodological boundaries of traditional humanistic scholarship. Their work
has been speculative, discursive, and rhetorical. They have drawn on empirical
research but have not, for the most part, adopted empirical methods. Instead,
they have used Darwinian theory as a source of
theoretical and interpretive concepts. In respect to method, then, their work
is similar to that of old-fashioned Freudians and Marxists.[1]
Aims and Designs
The study described in
this book is designed to help bridge the gap between empirical methods and
literary research. Building on findings in the evolutionary human sciences, we
aimed to (1) construct a basic working model of human nature, incorporating motives,
emotions, features of personality, and preferences in marital partners; (2) use
that model to analyze some specific body of literary texts and the responses of
readers to those texts, and (3) produce data—information that could be
quantified and used to test specific hypotheses about those texts.
Evolutionary social scientists
are still in the process of constructing a full and adequate model of human
nature. They know much already about how human reproductive behavior and human
sociality fit into the larger pattern of human evolution. They still have much
to learn, though, about the ways literature and the other arts enter into human
nature. In building a model of human nature, we draw on our knowledge of the
literary imagination, integrate that knowledge with evolutionary theories of
culture, and produce data that enable us to draw conclusions on an issue of
broad significance for both literary study and the evolutionary human sciences:
the adaptive function of literature and the other arts.[2]
In order to make advances
in knowledge, it is necessary to choose some particular subject. Genetics is a
basic science that applies to all organisms, but geneticists first got an
empirical fix on their subject by focusing minutely, with Mendel, on peas, and,
with Morgan, on fruit flies. In place of peas and flies, we have taken as our
subject British novels of the longer nineteenth century (Austen to Forster). We
wish to make it clear, though, that our aims extend well beyond the specialist
fields of British novels, the nineteenth century, British literature, narrative
fiction, or even literary scholarship generally. We wish to engage the
attention of literary scholars in all fields, and we also wish to engage the
attention of social scientists. We wish to persuade literary scholars that
empirical methods offer rich opportunities for the advancement of knowledge
about literature, and we wish to persuade social scientists that the
quantitative study of literature can shed important light on fundamental
questions of human psychology and human social interaction. Our own research
team combines these two prospective audiences. Two of us (Carroll and
Gottschall) have been trained primarily as literary scholars, and two of us (Johnson
and Kruger) primarily as social scientists. We have pooled our various forms of
expertise and along the way have acquired some familiarity with the knowledge
and skills in which our colleagues can claim special expertise.
We created an on-line questionnaire,
listed about 2,000 characters from 201 canonical British novels of the
nineteenth and early twentieth centuries, and asked respondents to select
individual characters and answer questions about each character selected. We
identified potential research participants by scanning lists of faculty in
hundreds of English departments worldwide and selecting specialists in
nineteenth-century British literature, especially scholars of the novel. We
also sent invitations to multiple listservs dedicated to the discussion of
Victorian literature or specific authors or groups of authors in our study. Approximately
519 respondents completed a total of 1,470 protocols on 435 characters. (A copy
of the questionnaire used in the study can be accessed at the following URL:
http://www-personal.umich.edu/~kruger/carroll-survey.html. The form is no
longer active and will not be used to collect data.) To ensure that we had
enough data on at least one novel to give a detailed account of that novel, we
set up a separate website for Thomas Hardy’s The Mayor of Casterbridge—the
subject of chapter eight. For the Mayor website, 85 respondents
completed 124 protocols on six characters.
Most of the respondents
were well educated, and many were professional literary scholars. They
evidently had excellent familiarity with the novels from which they selected
characters for coding, and they displayed high levels of competence as readers.
Chapter five provides more detailed information about the respondent pool and
offers statistical evidence that the respondents provided high-quality,
reliable data.
The questionnaire contains
three sets of categories. One set consists in attributes of characters: age,
attractiveness, motives, the criteria of mate selection, and personality. (The
sex of the characters was a given.) A second set consists in readers’ emotional
responses to characters. We listed ten possible emotional responses and also
asked whether the reader wished the character to succeed in achieving his or
her goals. The third set consists in four possible “agonistic” role
assignments: (1) protagonists, (2) friends and associates of protagonists, (3)
antagonists, and (4) friends and associates of antagonists. Respondents were
free to fill out questionnaires on any individual characters from the list. For
each character selected, respondents assigned scores on each category and also
assigned the character to one of the four possible agonistic roles. Alternatively,
the respondents could decline to assign a character to a role.
Sex, Gender, and Agonistic Structure
Our analytic scheme
focuses on relations among two main elements in the organization of characters:
(1) sex and gender, and (2) agonistic structure. By sex, we mean biologically
based differences of motivation between males and females. By gender, we mean
the way biological differences of sex are organized into the social roles that
vary widely from culture to culture. By “agonistic structure,” we mean the polarized
contrasts between protagonists and their associates on the one side, and
antagonists and their associates on the other. We refer to protagonists and
antagonists together as “major” characters and to their associates as “minor”
characters. Following popular usage, we refer to protagonists and their
associates as “good” characters and to antagonists and their associates as
“bad” characters. In adopting the common terms “good” and “bad,” we intend no
pre-emptive moral judgments. As it happens, though, the data indicate that good
and bad characters are heavily inflected with morally relevant traits. Popular
usage evidently has its reasons.
Contrasts between male and
female characters are designated differences of “Sex,” contrasts between major
and minor characters differences of “Salience,” and contrasts between good and
bad characters differences of “
Sex is a biological
category massively confirmed in empirical research—with differences in anatomy,
hormonal influences, and even neurological structures.[3]
Agonistic structure, in contrast, is a literary
category that has never been empirically investigated on any large scale. In
electing to investigate agonistic structure empirically, we did not presuppose
the validity of the concept. Instead, we posed the very existence of agonistic
structure as a question for research. Does agonistic structure actually exist
in these novels? Can characters be designated major and minor, good and bad, in
any intelligible and consistent way? Alternatively, is agonistic structure just
a popular misconception, a false and shallow notion reflecting the prejudices
of naïve readers? Are characters in fiction too complex and morally mixed to be
meaningfully characterized as “good” and “bad”? Are criteria determining
Salience too fine or too variable to make strong distinctions between “major”
and “minor” characters? Though such terms are often used in a loose,
conventional way among literary critics, could it be that they have little
relevance to the deeper structures of meaning in literary texts? When critics
speak of major and minor characters, or protagonists and antagonists, do those
terms isolate central organizing principles in the novels? Or, to the contrary,
on closer inspection, do the terms simply disintegrate? If the terms can in
fact be meaningfully applied to individual novels, how much does agonistic
structure vary from novel to novel? Does agonistic structure produce any
patterns on a scale sufficiently large to subsume differences among individual
authors and individual works—for instance, differences of sex, political orientation,
subject matter, and genre?
Researchers seldom conduct
experiments merely to prove some concept wrong. More often, they wish to test
concepts they consider potentially robust and fertile. They want to find out
how well the concepts hold up under analysis and how far they can be used in
reaching conclusions of real significance. The question as to whether agonistic
structure exists is a real empirical question. The answer could have gone
either way. To be frank, because of local glitches in statistical analysis,
there were unpleasant moments when we had a dismayed feeling the answer would
go contrary to our expectations. What we expected was that agonistic structure
would prove a robust and fertile concept. We predicted that each character set
would stand out sharply and distinctly from the others. Further, we anticipated
that the differences among the character sets, and especially the differences
between protagonists and antagonists, would illuminate the cultural ethos
embodied in the novels. We expected that protagonists would exemplify the
values that authors favored and that they thought their readers would favour.
Antagonists would exemplify values of which authors and readers disapprove. If
that should be the case, we wanted to see just how far agonistic structure
would take us in understanding how the novels organize the elements of human
nature.
To explore sex and gender
in the novels, we compared male and female characters, the responses of male
and female readers, and the representation of male and female characters by
male and female authors. We had no real doubt that differences of sex would
appear in characters. About differences in the responses of male and female
readers and differences in characters by male and female authors, though, we made
no specific predictions. For those two topics, we were willing to perform the
statistical comparisons and see what conclusions we could draw from the
results. We found, rather surprisingly (to us) that male and female respondents
do not significantly differ in their responses to the novels, neither in the
way they identify the attributes of characters nor even (more surprisingly) in
their emotional responses to characters. That finding serves as one form of
evidence in our argument for authorial control of meaning, which in turn serves
as evidence that literary meaning is determinate and quantifiable (chapter
five). The comparison of male and female characters by male and female
novelists has a chapter to itself (chapter four). Those findings provide a context
for our analysis of the way Jane Austen feminizes protagonistic males (chapter
eight).
We were interested in sex
and gender, and we were interested in agonistic structure, but we were
especially interested in the interactions between sex, gender, and agonistic
structure. How are the basic elements of an evolved sexual psychology organized
in the gendered social roles of males and females in this period? How does that
organization enter into the tonal and thematic structure of the novels? In the
past thirty or forty years, more criticism has been written from a feminist
perspective than from any other. That perspective typically emphasizes male
oppression and rebellious self-assertion in female protagonists. How closely
does that perspective align with quantitative findings on the relations between
male and female characters? Would those relations differ in male and female
novelists? How would relations between all male and all female characters
compare with relations between male and female protagonists and antagonists?
Would, for example, female protagonists exemplify characteristics less typical
of domesticated female roles? Would female protagonists display more
self-assertion? More generally, how would the normative cultural values
invested in agonistic structure interact with basic, universal dispositions of
an evolved sexual psychology? Answers to those questions, and others like them,
form the chief topics in chapters four and six.
Agonistic structure is
deeply embedded in the human imagination. It influences most phases of our
imaginative life—religion, philosophy, history, political ideology, workplace
gossip, video games, sports, movies. As an unexpected benefit—one of the
rewards of a research design robust enough to capture universal features of the
human imagination—analyzing agonistic structure in the novels also provides
concepts useful for analyzing the imaginative character of feminist literary
theory. Feminist theory is troubled by a nagging, unresolvable conflict between
“social constructivism” and “essentialism”—the contrasting ideas that sexual
identity is an arbitrary social convention and that it is an irreducible,
transcendent category. We identify the elements of truth in both constructivism
and essentialism, reconcile them, and suggest a more consistent and
comprehensive framework for analyzing gender in both life and fiction (chapter
six).
Big Ambitions, Great Expectations
One of our chief aims in
this study is simply to demonstrate that categories from evolutionary
psychology can be used to capture important structures of meaning in novels.
That finding would lend strong support to the argument that literary meaning
can be delimited, quantified, and located within a consilient, biological
understanding of human nature. As we define it, “meaning” consists, in the
first place, in the psychological states—ideas and feelings—that literature
produces in the minds of readers. In their total organization with respect to
any given novel, we refer to these aggregated psychological states as “thematic
and tonal structures.” We have not included stylistic organization as one of
the categories in this study but are confident that style, like theme and tone,
has objectively discernible attributes that produce distinct psychological
states in readers. Moreover, we are confident that formal organization
interacts with thematic and tonal organization in producing determinate
structures of meaning.
When we say that meaning
can be precisely and objectively determined, we mean that the psychological states
produced in readers can be measured in the same way that one can determine the
dimensions of, say, a room. One can specify the dimensions (height, width, and
depth) and measure them. In this study, we specify the dimensions of tonal and
thematic structure, deriving them from a model of human nature, and measure
them, both in the aggregate (441 characters, sorted into sets and averaged in
each set) and in particular cases (441 characters, each taken individually). We
check and re-check these measurements in multiple ways: assessing the degree of
correspondence among multiple coders (chapter five); comparing redundant
measures of the same dimensions—for example, emotional responses, role
assignments, and stipulations as to whether respondents wish a character to
succeed and whether they think that character’s success is a main feature of a
given novel; correlating scores on the attributes of characters in several
different categories—age, attractiveness, motives, mate selection, and
personality; correlating scores on these attributes with scores on emotional
responses; and correlating both attributes and emotional responses with
agonistic role assignment (chapters one through four).
All of these efforts in
measurement are animated by beliefs similar to those that have produced the
modern, scientific understanding of human personality. Describing the
historical development of this understanding, Daniel Nettle observes that “most
of the work in scientific psychology consists in trying to come up with good measures
of things, and showing that they are good measures. Indeed, a concern with
measurement is precisely what distinguishes ‘academically respectable’
psychology from psychology of other kinds” (Personality 17). The purpose
of developing measures for personality is to bring it “within the fold of
scientifically studiable entities.” Our ambition, most simply, is to bring
literary meaning within the fold of scientifically studiable entities.
Measurement anchors our
claim for adopting an empirical methodology, but one can measure things without
achieving very much in the way of real knowledge. Witness the widely
acknowledged futility of so much of the work in sociology and behaviorist
psychology. To make real advances in knowledge, one must know what is worth measuring.
To measure a room presents a technical challenge, but not much of a conceptual
challenge. Height, width, and depth have the kind of a priori transparency that
wins spontaneous unanimous consent. They are axiomatic. We do not anticipate
achieving that level of confidence in selecting the dimensions we choose to
measure. We do, however, anticipate achieving levels comparable to those in
personality psychology. And indeed, the five factors of personality are
themselves important dimensions in our analytic design. In isolating basic
motives, basic emotions, and the criteria for selecting mates, we aim at
capturing dimensions more or less equivalent, in scope and reliability, to
the five factors of personality.
We reserve our major
discussion of determinate meaning for chapter five, after we have reported the
results of the study. But since this issue is of broad interest for literary
theorists, and since our claims are likely to arouse immediate concerns in the
minds of readers, we shall pause here long enough to give a preliminary account
of our argument on meaning. Insofar as meaning consists in the psychological
states created by literature in the minds of readers, it can be measured. If
qualified readers agree in affirming simple and measurable attributes of
characters in a novel, that shared understanding can be cited as evidence
supporting the actual existence of those attributes. If readers display common,
measurable emotional responses to those attributes, these shared responses can
be cited as evidence for a determinate and largely intentional structure of
meaning in that novel. Novelists create attributes in characters in order to
produce predictable responses in readers. The total pattern of attributes and
responses, coordinating with agonistic role assignments, produces thematic and
tonal structures that constitute the common understanding of the novel—the
experience of “the common reader.” The
common experience of the common reader is thus in itself not merely a
scientifically ascertainable fact; it provides a basis for affirming the objective
reality of literary meaning.
What of history? The
dominant mode of thinking in contemporary literary study attributes all
“meaning” to “culture,” to supposedly unique moments of historically
conditioned thoughts, feelings, and values. The notion that unique cultural
configurations are autonomous sources of meaning was decisively refuted in 1991 in Donald
Brown’s book Human Universals. Brown is an evolutionary anthropologist
who delineated dozens of features of social organization, cognition, and
imaginative activity that appear in all known cultures. Any universal feature
of culture is almost certainly strongly conditioned by genetically transmitted
dispositions—by the universal, biologically based characteristics that
distinguish human beings as a species. Attributing all meaning to culture
presupposes a “blank slate” model of the mind, and it necessarily operates as a
tautology: “thoughts, beliefs, feelings, and values are all produced by
culture; culture consists in thoughts, beliefs, feelings, and values. Thoughts,
feelings, beliefs and values thus cause themselves.” The evolutionary human
sciences offer a simple and sensible alternative to this vicious cultural
circle. As Brown’s work confirms, there are no truly unique cultural
configurations; there are only variations on the culture of “the universal
people." In the degree to which human groups are genetically identical,
those variations can be produced only by differences in environmental conditions:
ecology, technology, modes of production and distribution, forms of social
organization, and the available traditions of imaginative culture—religion,
philosophy, ideology, and the arts. All that together is “history.”
History is not a transcendent, irreducible, ultimate cause, an unmoved mover or
primum mobile. It is merely the sequence of cultural configurations
produced by the interaction of genetically transmitted human dispositions and
environmental conditions. Those conditions include but are not limited to
available traditions of imaginative culture. Moreover, all available traditions
of imaginative culture have themselves been produced by interactions between
genetically mediated dispositions and environmental conditions.
With a sufficiently robust
and precise set of analytic categories from an evolutionary understanding of
human nature, and with a sufficiently comprehensive understanding of the
environmental conditions with which the elements of human nature interact, it
is hypothetically possible to give a satisfactory causal account of any
specific historical configuration of cultural values. And not just
hypothetically possible. Evolutionary cultural historians are already making
real progress in offering biologically informed interpretations of major phases
of historical consciousness (Smail) and in developing fundamental principles
for analyzing large-scale historical movements such as the rise and fall of
empires (Turchin). One of our contributions in this book is to provide a set of analytic concepts that can be
used for delineating the ethos of a specific culture, as manifested in a large
body of novels, for locating that cultural ethos within the larger context of
human evolutionary history, and for assessing the specific configuration
of thematic and tonal aspects of meaning
in any given novel. We embed the analysis of thematic and tonal structures in
individual novels within a specific cultural ethos and analyze that ethos as a
specific organization of universal features of human nature.
In addition to
demonstrating the determinacy of meaning, we hoped, rather vaguely at first,
that the findings in this study would shed light on whether novels fulfill
basic needs of the adapted mind—and if they do, how they do it. The most
important controversy in evolutionary literary study is whether any of the arts
fulfill any adaptive functions. We hoped our findings on the organization of
characters could lead us to significant conclusions on whether these particular
literary works fulfill some adaptive function. We were not sure what
that function might be. “Literature” of course depends on literacy, a very
recent acquisition in human evolutionary history, so recent that it cannot
plausibly be considered an adaptation. But the oral antecedents of literature
are part of the basic profile of “human nature.” People in all non-literate
cultures use language, tell stories, and play with words in creative and
evocative ways (Brown, Human Universals). Written language is just a
cultural technology extending those universal human aptitudes. If oral forms of
agonistic structure could plausibly be considered mechanisms that fulfill some
adaptive function, the same mechanism would fulfill an adaptive function in
written forms.
We anticipated that
protagonists would form cooperative communities and that antagonists would
mainly seek power in selfish ways, but we did not at first lodge those
predictions within any larger explanatory framework. Like most readers, we felt
intuitively that it must be so, that protagonists and their friends are nice
guys and that antagonists and their associates—antagonists tend not to have
“friends”—are not. Using categories from evolutionary psychology, we were able
to give precise empirical content to the terms “good” and “bad,” but that was
where we stopped, at first. In the course of analyzing the data, we became
aware that the ethos reflected in the novels replicates the egalitarian ethos
of hunter-gatherers, who stigmatize and suppress status-seeking in potentially
dominant individuals (Boehm). Suppressing or muting dominance within a social group enhances group
solidarity and organizes the group psychologically for cooperative endeavor. The
parallelism between agonistic structure in the novels and the egalitarian ethos
of hunter-gatherers has a clear implication for the adaptive function of
agonistic structure. If suppressing dominance in hunter-gatherers fulfills an adaptive social
function, and if agonistic structure in the novels stigmatizes dominance
behavior, our study would lend support to the hypothesis that agonistic
structure fulfills an adaptive social function. Agonistic structure would serve
as a medium for affirming an egalitarian ethos on a large cultural scale. It
would be a mechanism supporting the formation of cooperative social groups.
This issue forms the chief burden of our argument in chapter seven.
Human Nature
The central concept in the
evolutionary human sciences and in evolutionary literary study is “human
nature,” the genetically transmitted dispositions derived from human
evolutionary history. Evolutionists presuppose that human nature informs and
constrains the behavior of individuals and social groups. Many—ourselves
included—also believe that human nature shapes the products of the human
imagination. The concept of human nature provides the categories we use for
analyzing the attributes of characters and the emotional responses of readers.
An understanding of the evolved social and political dynamics of human nature
also frames the conclusions we draw on the adaptive function of agonistic
structure. It would be well, then, to say something here about what we mean by
human nature.
Until fairly recently in
literary history, most writers and literary theorists presupposed that human
nature was their subject and their central point of reference. Dryden following
Horace, who follows others, offers a representative formulation. In “Of
Dramatic Poesy,” Dryden’s spokesman Lisideius defines a play as “a just and
lively image of human nature, representing its passions and humours, and the
changes of fortune to which it is subject; for the delight and instruction of
mankind” (25).[4]
The understanding of human nature in literature is the most articulate form of
what evolutionists call “folk psychology” (Boyer;
Postmodernists have put
all such ideas of human nature out of play, thus cutting themselves off from
the accumulated stores of wisdom in folk psychology. Evolutionary social
scientists, fortunately, have taken a different path. While literary theorists
were immersing themselves in speculative theoretical systems such as
phenomenology, psychoanalysis, deconstruction, and Marxism, the evolutionists
were gradually developing an empirically based model of human nature, including
childhood development, family dynamics, sexual relations, social dynamics, and
cognition.
In the early days of
sociobiology, through the 1980s, evolutionary theorists of human nature
concentrated on “inclusive fitness”—passing on genes through offspring or other
kin.[5] In
the 1990s, “evolutionary psychologists” distinguished themselves from
sociobiologists by emphasizing “proximate mechanisms” that mediate reproductive
success, but they still did not produce a whole, usable model of human nature.
Instead, they produced open-ended and unorganized lists of “modules,” dedicated
bits of neural machinery that were supposed to have solved specific adaptive
problems in ancestral environments. Modules were postulated for sense
perceptions, various forms of subsistence activity, natural history, selecting
mates, detecting cheaters, emotions, avoiding predators, “and so on” (Carroll, Literary
Darwinism 106-07). As a complement to lists of modules in evolutionary
psychology, Donald Brown offered a list of “human universals,” that is, practices
found in all known cultures and thus presumably constrained by the evolved and
genetically transmitted features of human nature. Human universals and
domain-specificity have remained important concepts in human evolutionary
theory, but over the past decade or so behavioral ecologists and developmental
psychologists have finally made available the crucial idea that had been
missing from these lists—the idea of a total systemic organization in human
nature. A scholar or scientist adopting a systemic perspective envisions all
the parts of a system as functionally interactive. Variation in one part
affects relations among all the parts. As a concept of structure, this idea is
essentially the same as that of “organic unity” espoused by Samuel Taylor
Coleridge and other Romantics.
The most comprehensive
concepts for the systemic organization of the parts of human nature derive from
“human life history theory.” All species have a “life history,” a
species-typical pattern for birth, growth, reproduction, social relations (if
the species is social), and death. For
each species, the pattern of life history forms a reproductive cycle. In the case of humans, that cycle centers on parents, children, and
the social group. Successful parental care produces children capable, when
grown, of forming adult pair bonds, becoming functioning members of a
community, and caring for children of their own. “Human nature” is the set of
species-typical characteristics regulated by the human reproductive cycle. This concept of human nature assimilates the
sociobiological insight into the “ultimate” importance of inclusive fitness as
a regulative principle, and it allocates proximal mechanisms a functional place
within the human life cycle. Early models of “the adapted mind,” concentrating
too exclusively on “modularity,” had excluded the idea of flexible general
intelligence. Using human life history as a systemic concept enables
evolutionists to integrate domain specificity with a flexible general
intelligence (Geary, The Origin; Kaplan and Gangestad 122).[6]
Human
beings have a life history that is similar in some ways to that of their
nearest relatives the chimpanzees (A. Buss; de Waal), but humans also have unique
species characteristics deriving from their larger brains and more highly
developed forms of social organization. Unlike chimpanzees and most other
mammals, humans display pair-bonded male-female parenting; and unlike all other
animals, they combine pair bonding with complex social organizations involving
cooperative groups of males (Flinn and Ward; Geary and Flinn). Humans take
longer to grow up, allowing time for their brains to mature and their social
skills to develop. And finally, for humans culture has an importance it does
not have in other species. Other
species have adaptations for cooperation in social groups with specialized
functions and status hierarchies. Other animals engage in play, produce
technology, and share information. Humans alone produce imaginative artifacts
designed to provide aesthetic pleasures, evoke
subjective sensations, express emotions, depict nature or human experience, or
delineate through symbols the salient features of their experience. The
genetically mediated dispositions of human nature—survival, mating, kinship,
friendship, dominance, cooperative group endeavor, and inter-group
competition—have evolved in a reciprocally causal relationship with the
cognitive and behavioral dispositions for producing and consuming the arts.
That causal interdependence is part of the evolutionary process evolutionists
denote as “gene-culture co-evolution”—a topic to which we give more
concentrated attention in chapter seven.[7]
Symbolic thinking is an
indispensable mechanism in the process of “group-selection” that has fuelled
much of human social evolution. Humans
identify themselves symbolically as members of groups larger than the bands
that mark the size limit for groups of chimpanzees. In humans, “tribal
instincts” are a basic part of
gene-culture co-evolution (Boehm; Richerson and Boyd, Not by Genes
230-35). By developing a capacity to subordinate individual dominance to the
cooperative endeavour of a social group, humans set off an evolutionary arms
race that has resulted in ever larger and more complex forms of social
organization.[8]
Human Nature as a Basis for Shared Understanding
The
culture in which an author writes provides a proximate framework of shared
understanding between the author and his or her projected audience, but every
specific cultural formation consists in a particular organization of the
dispositions of human nature, and those dispositions form the broadest and
deepest framework of shared understanding. Many of the authors in this study
make overt and explicit appeals to “human nature.” By delineating the features
in the folk concept of human nature, we can reconstitute the understanding
authors share with readers. That shared understanding includes intuitions about
persons as agents with goals, differences of sex and age, basic human motives,
basic emotions, and the features of personality.
Traditional
humanists do not often overtly repudiate the idea of human nature, but they do
not typically seek causal explanations in evolutionary theory, either. In the
thematic reductions of humanist critics, characters typically appear as allegorical
embodiments of humanist norms—metaphysical, ethical, political, psychological,
or aesthetic. In the thematic reductions of postmodern critics, characters
appear as allegorical embodiments of the terms within the source theories that
produce the standard postmodern blend—most importantly, deconstruction,
feminism, psychoanalysis, and Marxism. In their postmodern form, all these
component theories emphasize the exclusively cultural character of symbolic
constructs. “Nature” and “human nature,” in this conception, are themselves
cultural artifacts. Because they are contained and produced by culture, they
can exercise no constraining force on culture. Hence Fredric Jameson’s dictum
that “postmodernism is what you have when the modernization process is complete
and nature is gone for good” (Postmodernism ix). From the postmodern
perspective, any appeal to “human nature” would necessarily appear as a
delusory reification of a specific cultural formation. By self-consciously
distancing itself from the folk understanding of human nature, postmodern
criticism loses touch both with biological reality and with the imaginative
structures that authors share with their projected audience. In both the
biological and folk understanding, there is a world outside the text. From an
evolutionary perspective, the human senses and the human mind have access to
reality because they have evolved in adaptive relation to a physical and social
environment about which the organism urgently needs to acquire information
(Lorenz). An evolutionary approach shares with the humanist a respect for the
common understanding, and it shares with the postmodern a drive to explicit
theoretical reduction. From an evolutionary perspective, folk perceptions offer
insight into important features of human nature, and evolutionary theory makes
it possible to situate those features within broader biological processes that
encompass humans and all other living organisms.
The questions
in the questionnaire we used to collect data are couched in the common language
and pitched at the level of common understanding, but they are also formulated
within the framework of an evolutionary model of human nature. The questions
are thus situated at the point at which the evolutionary model converges with
the common understanding. The questions register the common understanding, quantify
it, and locate it within the context of empirical social science.
Quantification enables us to give an objective, formal analysis of the common
understanding and to assess statistically the structural relations among its
elements. Rendering literary knowledge as data makes it possible to treat it
empirically, to make predictions, test hypotheses, link specific findings with
other empirical studies in adjacent areas of social science, and, consequently,
to develop knowledge in a way that could be cumulative and progressive. This
kind of knowledge overlaps with common observation and with the formulations of
literary critics, but as empirical data that enables the precise, quantitative
testing of hypotheses, it is epistemologically of a different order.
Measures of Success for this Research Design
In using an evolutionary
model of human nature for the analysis of these novels, we have presented
ourselves with a dual challenge. Is this model robust enough to account for
broad structures of meaning in a large body of texts? Is it also fine-grained
enough to give us interpretive access to specific literary texts? The lowest
standard for success in meeting these two challenges would be to replicate the
common understanding. Satisfying that standard would demonstrate that major features of literary
meaning can be effectively reduced to simple categories grounded in an
evolutionary understanding of human nature. A higher standard would be to discover
something new and important. Distinguishing these two standards is useful for
descriptive purposes but should not obscure the fact that the two levels
interact. Merely replicating the common understanding produces evidence
supporting a controversial issue at the very heart of literary study—the
question as to whether meaning is determinate, whether it can be delimited,
quantified, and confirmed empirically (chapter five).
Issues involving broad
structures of meaning include the existence and nature of agonistic structure,
relations between Sex and
The
same standards apply to the more fine-grained analysis of individual texts.
First, are the categories used in this study sufficient to replicate the
findings of ordinary discursive criticism? If ordinary discursive criticism can
be subsumed within an evolutionary model of human nature, we could conclude
that literature occupies no mysteriously qualitative realm set apart from the
world knowable by science. That implication contravenes an idea, common in the
humanities, that literary meaning is illimitably complex and contains
irreducible elements of the qualitatively unique.[9] Offering
evidence that contravenes humanistic dualism would in itself be worth the
effort of inquiry. But merely negative findings do not provide the highest
satisfactions of research. Scholars and scientists alike are ambitious of
discovery. Can our concepts and methods generate new insights into familiar
texts? We think they can. The proof, of course, is in the readings.
Putting the Categories to the Test
In chapter one, mercifully
short, we explain how to read the scores on characters and character sets. In chapter two, we delineate a single basic
pattern across a large body of novels: agonistic structure differentiated by
sex. In chapter three, we reverse analytic direction, using that single basic
pattern as a base line against which to delineate the individual structures of
meaning in a diverse array of specific novels. We interpret the scores of
characters who closely exemplify the norms or averages in the data set as a
whole and also the scores of characters who cut across the grain, putting heavy
pressure on the normative patterns that guide the expectations of readers.
The characters selected for detailed attention in chapter three are
drawn chiefly from a pool of 48 characters who received seven or more codings
each (appendix three). Averaging
scores from multiple codings minimizes the effect of any idiosyncratic
individual codings and thus provides greater confidence about the reliability
of the scores. Chapter five provides much more detail on measuring reliability
in scores.
In chapter four, agonistic structure provides a
framework for comparing male and female characters by male and female authors. Chapters five, six, and seven are each devoted to one main large
issue raised by the analysis of agonistic structure: the determinacy of
meaning, sexual politics, and the adaptive function of literature and the other
arts. Chapters eight and nine offer case studies of the novels for which we
have the most abundant data. In the conclusion, we compare our governing ideas
and methods to those of other schools.
Jane Austen received more
codings than any other author, enough so that we can identify the thematic and
tonal patterns across her whole body of work, displaying those patterns in
graphs—hence the title of the book. We offer quantitative evidence on a
disputed issue about Austen’s ideological orientation, explain how her
depiction of male sexual identity modifies the usual form of romantic comedy,
and describe how she integrates an egalitarian ethos with the satisfactions of
rank and wealth. Taking account of how all these features interact produces a
new perspective on the kind of emotional appeal Austen has for her readers. In
the critique of Austen, we offer an advance on the critical consensus. For The
Mayor of Casterbridge, no such consensus exists. Analyzing agonistic
structure in Mayor produces a new solution for an interpretive problem
that has eluded traditional critical methods for more than a century.
The Scope of Our Claims
On the basis of the data
collected through the questionnaire, we make three large claims: (1) that the
novels in this study contain determinate structures of meaning that can be
captured using the categories in our research design (chapter five); (2) that
differences of Valence are much more structurally prominent than differences of
Sex (chapter six); and (3) that agonistic structure in these novels fulfills an
adaptive social function (chapter seven). Assuming we can make the case
for all three of those claims, how far can we generalize from those conclusions
to all literature, in every period and every culture? Logically, it is possible
that no other literary texts anywhere in the world contain determinate
meanings, display differences between good and bad characters more prominent
than differences between male and female characters, or fulfill any adaptive
function at all. Hypothetically possible, but not very likely. If our arguments
hold good for this body of texts, they demonstrate that determinate meaning is
at least possible, that in at least one body of classic texts
In arguing that agonistic
structure in these novels fulfills an adaptive social function, we do not
suppose that we have isolated the sole adaptive function of all literature.
Quite the contrary. Along with other evolutionary literary theorists, we
strongly suspect that literature fulfills other functions.[10]
When we take up the issue of adaptive function in chapter seven, we locate our
arguments in the context of the controversy over whether the arts fulfill any adaptive function at all. We argue that
the social dynamics animating these novels derive from ancient, basic features
of human nature. Such features would in all likelihood appear in some fictional
narratives in most or all cultures. We would of course be interested to know
whether the kind of agonistic structure we identify in these novels is in fact
a human universal. If it is a human universal, we would also be interested to
know how it varies in form in different cultural ecologies. (Marriage is a
human universal but varies in form from culture to culture. We might expect
agonistic structure, like marriage, to vary in form.) These questions would
make good topics of research for other studies. Until those studies are
conducted, though, the topics are only a matter for theoretical speculation.
For this current study, we can positively affirm only the conclusions we think
our data allow us to draw. Hence the limiting terms in our subtitle:
paleolithic politics in British novels of the nineteenth century.
[1] On the two cultures, see Boghossian; Fromm, “My Science Wars,”
“Science Wars and Beyond,” The Nature; Gottschall, “The Tree”; Gross and
Levitt; Gross, Levitt, and Lewis; Koertge; Leavis, “Two Cultures?”; Parsons; Slingerland;
Snow; Sokal; Sokal and Bricmont; Weinberg, Dreams, Facing Up; E.
Wilson. For a synoptic historical account of modern literary theory, see Abrams,
“The Transformation.” For surveys
of contributions to evolutionary literary study, see Carroll, “An Evolutionary
Paradigm,” “Evolutionary Approaches.” For arguments on using empirical methods
to renovate literary study, see Gottschall, Literature, “Quantitative
Literary Study.” For examples of evolutionary literary study by both humanists
and scientists, see Boyd, Carroll, and Gottschall; Gottschall and
[2] On the adaptive function of the arts, see Boyd, “Evolutionary
Theories” 147-76, On the Origin; Carroll, “An Evolutionary Paradigm”
119-28, Literary Darwinism 63-68,
“Rejoinder” 349-68; Dissanayake; Dutton; Pinker, How the Mind Works
521-44, The Blank Slate 400-20; Salmon and Symons, “Slash Fiction”;
Tooby and Cosmides, “Does Beauty Build?”; E Wilson ch. 10.
[3] For empirical inquiries into sex and gender, see Baron-Cohen, The
Essential Difference; Baron-Cohen, Lutchmaya, and Knickmeyer; Barrett,
Dunbar, and Lycett; D. Buss, The Evolution; Costa, Terraciano, and
McCrae; Dabbs; Ellis and Symons; Gangestad; Goldberg; Gottschall, “The
Heroine”; Geary, Male; Hrdy; Jones; Kimura; Lippa; Schmitt; Panksepp
147-48; Salmon and Symons, “Slash Fiction,” Warrior Lovers; Symons, The
Evolution; Trivers; Vandermassen.
[4] For references to other such examples, see
Carroll, Evolution 170; Pinker, The Blank Slate 404-20.
[5] For examples of
“sociobiology,” see Alexander, The Biology;
Betzig; Chagnon; Chagnon and Irons; Dawkins, The Selfish Gene;
[6] For critiques of “narrow” or “orthodox”
evolutionary psychology, see Barrett, Dunbar, and Lycett 8-21; Carroll, “An
Evolutionary Paradigm” 122-24, “The Human,” Literary Darwinism 190-99;
Dunbar and Barrett, “Evolutionary Psychology”; Griffiths 106-36; Hill; Laland;
McDonald, “A Perspective”; Mithen; Panksepp; Panksepp and Panksepp; Sterelny.
For evolutionary accounts of human
cognitive architecture broader than that in orthodox evolutionary psychology,
see Geary, “Motivation to Control” and The Origin; Sterelny. On the
developmental plasticity of human cognitive architecture, see Deacon; Geary; Panksepp;
Plotkin; Sterelny.
[7] On human life
history theory, see Carroll, “An Evolutionary Paradigm” 111-15; Hill; Kaplan
and Gangestad, “Life History Theory,” “Optimality Approaches”; Kaplan, Hill,
Lancaster, and Hurtado; Lancaster and Kaplan; Low; Lummaa. On the evolution of
the human family—childhood, mating, and parenting—see Bjorklund and Pellegrini;
D. M. Buss, The Evolution; Deacon; Flinn, Geary, and Ward; Flinn and
Ward; Geary, “Evolution,” Male; Geary and Flinn; Schmitt. On the
evolution of social cognition in humans, see Baron-Cohen, “The Empathizing
System”; Budiansky; A. Buss; Darwin, The Descent; Focquaert and Platek;
Guise et al.; Hauser; Lewis; Paulhus and John; Premack and Premack; V. Stone;
Tomasello et al.; Zunshine. On the unique human capacities for culture, see
Baumeister; Carroll, “An Evolutionary Paradigm,” “Rejoinder”; Deacon;
Dissanayake; Mithen; Panksepp and Panksepp; Wade; E. Wilson. For contributions
to the theory of gene-culture co-evolution, see Barrett, Dunbar, and
Lycett 351-83; Boyd and Richerson; Henrich and McElreath; Hill; Laland; Lumsden
and Wilson; McElreath and Henrich; Plotkin; Richerson and Boyd, Not by Genes,
“The Evolution”; Shennan; Smail; Sterelny; Tomasello et al.; D. Wilson, Evolution;
E. Wilson, Consilience.
[8] On the social dynamics of dominance and cooperation,
see Alexander, “Evolution”; Axelrod and Hamilton; Bingham; Boehm;
Cummins; Darwin, The Descent 1: 70-106; Eibl-Eibesfeldt,
“Us”; Flinn, Geary, and Ward; Kurland and Gaulin; Richerson and Boyd, Not by
Genes, “The Evolution”; Ridley; Salter; Smail; Sober and Wilson; Sterelny;
Turchin; Wade; D. Wilson, Darwin’s Cathedral, Evolution, “Human
Groups,” “the Role.” On the multiform linguistic devices that mediate polite behavior in
contemporary society—behavior designed to deflect the appearance of seeking
dominance—see Pinker, The Stuff 373-425; Salter 71-87. On the
now largely successful effort to resuscitate the idea of “group selection,” see
Boehm; Eibl-Eibesfeldt, “Us”; Richerson and Boyd, Not by Genes; Sober
and
[9] For
recent affirmations of this humanistic credo, see Goodheart; Seamon;
Spolsky, “The Centrality.”
[10] See for instance Boyd, “Evolutionary
Theories”; Carroll, “An Evolutionary Paradigm”; Dissanayake; Dutton; Salmon and
Symons, “Slash Fiction”; Scalise Sugiyama, “Food,” “Reverse-Engineering
Narrative”; Tooby and Cosmides, “Does Beauty Build?”