Patrick Colm Hogan, University of Connecticut
Nigel Fabb’s essay on “Arbitrary Innovations and Literary Universals” treats the degree to which individual authors are able to defy literary practices, including practices that appear to be universal. This issue has a wide range of implications for the study of literary universals, all of which cannot be treated in a single essay. The present article is complementary to Fabb’s and should be read in conjunction with it. Specifically, the following paragraphs consider two issues related to arbitrary innovations. The first issue is the social adoption of innovations. It is addressed explicitly by Fabb, but with a different focus. The second issue is which innovations might or might not violate universals, and potentially falsify them. This is more general and only implicit in Fabb’s essay.
As to the first issue, Fabb points to the difference between innovations that are generally adopted within a tradition and those that are not. For example, we might ask why Gerard Manley Hopkins’s metrical innovations appear to have had very little influence on subsequent versification. As Fabb points out, one answer is that their introduction did not flourish due to historical contingency. (Fabb mentions literary history in particular, but it seems clear that non-literary contingency—such as racist or sexist prejudice against the innovator—may enter as well.) Another answer is that they are somehow in tension with universal patterns (e.g., they are not consistent with linguistic processing tendencies). Similar points might be made about innovations that do have wide social impact. For example, some techniques of Hollywood cinema (e.g., the practices of “continuity editing”) might have caught on because they conformed to universal perceptual propensities (see Bordwell), while others might have succeeded due simply to the accidental success of certain films or the dominant position of particular studios. Moreover, as the example of Hollywood suggests, the point applies not only within, but also across cultural traditions.
The cross-traditional problem is highly significant. As I have discussed elsewhere (“Areal”), it is much more difficult to quarantine literary traditions than languages. Thus, it is often difficult to ensure that one’s research has the appropriate degree of genetic and areal diversity. This is important because claims of universality require specific sorts of support. One has evidence that a feature is universal only if three criteria apply. First, the feature must recur across traditions with significantly greater frequency than would be expected by chance. Second, those traditions must not have descended from a common source with that feature. Third, those traditions must not have influenced one another with regard to that feature. The problem is that it may be very easy for one literary tradition to be influenced by another even with limited contact, and it is not easy to find literary traditions that have had no contact whatsoever. Thus, we may face a fundamental problem with evidence regarding literary universals.
In response to this problem, I have proposed several measures. Of course, we should try to find genuinely genetically and areally distinct traditions. Nonetheless, in some cases, evidence for a universal may be gleaned from traditions that have had contact. That contact, however, should not involve hegemonic relations such that the literary practices of one tradition would be likely to displace the literary practices of another tradition for non-literary reasons (e.g., because one society colonized the other and imposed its literary tradition through the education system). Thus, one should in general not draw on modern, post-colonial literatures for evidence of universals. Moreover, one should take into account the extent of the contact and the salience of the features. For example, as to the latter, the mere fact of dramatic performance is highly salient, thus easily transmittable even with minimal contact. In contrast, certain sorts of background imagery or the use of techniques such as foreshadowing are not likely to be highly salient unless people are already sensitive to them for some reason. Thus, they should not be transmitted so readily. If there is minimal contact and a feature has low salience, then the recurrence of the feature is less likely to have been the result of contact.
Finally, even if we have reason to believe that some feature has been transmitted by contact, we may still wish to incorporate it into a theory of universals. This is because not everything transfers across traditions, especially not everything that has a low degree of salience. We may learn something about universal processes of literary creation or reception or universal functions of literature from the fact that one practice transfers and another does not. Consider, for example, some feature that recurs across different traditions, but that has evidently passed from one tradition to another. If that feature has high salience and the contact has been extensive, then the recurrence has little if any value for the study of universals. If it has moderate salience and the contact has been intermediate, then the recurrence may have some value. If it has low salience and the contact has been limited, we may decide that the recurrence has considerable value. (There are, of course, other possibilities as well.)
The second issue raised by arbitrary innovation concerns, again, what contradicts or falsifies a posited literary universal. Here we need to note that universals differ in their scope and in their conceptual kind. As to scope, some universals may apply to all literary works. I suspect that these are mostly trivial and in general derive from our definition of literature. Arbitrary innovations might seem to falsify a universal that applies to all literary works. However, they are more likely to challenge our definition of literature. To take a simplistic example, suppose we define literature by reference to fictionality. Do autobiographical poems show that literature is not universal? That does not seem to be a reasonable conclusion. Rather, autobiographical poems suggest that we should probably not define literature as fictional. We may of course wish to investigate cross-cultural properties of fictional works. But we simply isolate those as works of fiction, not as the entire body of literature.
More importantly, the scope of universal claims is more often a matter of common practices within a tradition or simply some works in a tradition. As to the former, most claims about poetic line length, for example, do not appear to concern the limits of line length for all poems; they rather concern the normal or default line lengths for a given tradition (such as iambic pentameter in English; see, for example, my The Mind 37-44). Arbitrary innovations would have to displace common practices in order to falsify such universals.
Probably the most frequent claims for universality bear on some works in the tradition. For example, heroic narratives appear in a wide range of traditions. However, not every work is heroic. Indeed, not every work is heroic or romantic or sacrificial or one of the other cross-cultural genres (on these genres, see my “Story”). The universal is that this genre recurs cross-culturally, not that it exhausts the possibilities for story formulation. It is difficult to see arbitrary innovations as violating universals of this sort.
Finally, we may distinguish two conceptual varieties of universal—prototypes, on the one hand, and rules or necessary and sufficient conditions, on the other. Some universals are formulated in terms of prototypes. For example, heroic plots do not have strictly necessary and sufficient conditions, but rather involve greater or lesser approximation to a standard case. It is difficult to see how arbitrary innovations could violate prototype-based claims of universality, though of course authors may arbitrarily innovate highly non-prototypical stories. Other universals are formulated as rules. For example, Kiparsky argues that alliteration “seems to be found as an obligatory formal element only in languages where the stress regularly falls on the same syllable in the word, which then must be the alliterating syllable” (9). This rule could in principle be violated by arbitrary innovations, though the innovation would have to become a standard practice in this case.
Future research in this area should continue to examine the kinds and degrees of areal contact and the nature of salience, both theoretically and through specific case studies. It should also consider whether or not there are other factors that might affect the ease with which literary features may be transmitted. Researchers might also address the parallel issue regarding historical development from a common source, taking up for example the degree to which particular features are or are not stable in historical development. It may be that there are specific cases in which a feature persists despite clear pressures for it to change. These may then be cases where genetic continuity could play a role in the isolation of universals.
Bordwell, David. “Convention, Construction, and Cinematic Vision.” In Post-Theory: Reconstructing Film Studies. Ed. David Bordwell and Noël Carroll. Madison, WI: University of Wisconsin Press, 1996, 87-107.
Hogan, Patrick Colm. “Areal Distinctness and Literature.” In The Cambridge Encyclopedia of the Language Sciences. Ed. Patrick Colm Hogan. Cambridge: Cambridge University Press, 2011, 105-106.
Hogan, Patrick Colm. The Mind and Its Stories: Narrative Universals and Human Emotion. Cambridge: Cambridge University Press, 2003.
Kiparsky, Paul. “The Role of Linguistics in a Theory of Poetry.” In Essays in Modern Stylistics. Ed. Donald C. Freeman. New York: Methuen, 1981, 9-23.
 In addition, some claims about lines may rely on a technical definition, such that the apparent “line” in a poem may not be the line as set out in the technical definition (for a discussion of issues in the segmentation of poetry, including lineation, see Fabb’s discussion of poetic sections). For example, writing conventions may leave lines unmarked, split them in two, or combine them (so that one “line” of writing is actually two lines by the technical definition). In these cases, innovations would have to affect line length in the technical sense of “line.”