Zipf’s Law, Style and the Literary – Puritan Style, Puritan Nature

(This is part 4 of my series on Zipf’s Law and literary style. Please see the last part, a disclaimer here)

Puritan Style, Puritan Nature

With my methodology thoroughly discussed, my first case study will be the effect of Puritan “plain style” as derived from (in my account) Ramist logic on the Zipfian structure of their writings. This puritan “plain style” though much discussed has no universal definition. The term itself does a decent job explaining the relevant aspects of this linguistic practice. Puritan writers would have been expected to eschew the metaphysical exhortations common to Anglican or Catholic writers in favor of more straightforward explications of the doctrines and usages embodied in God’s world.   Though Perry Miller notes that “considerable freedom was possible” in the actual process of writing, he also points out that the expectation of a simple style extended through all aspects of Puritan discourse. ¹ Miller argues that this stricture was pervasive enough to devalue the concept of genre, leading to a literary field where “the authors of that [non-sermon] literature, who in most cases were divines, did not conceive of themselves as writing in literary genres.” ² Indeed, literary genre took a back seat to the means and usages derived from the predominantly Ramist intellectual context that supported these Puritan writers. Ramism, a simplification of Aristotle’s system of logic promulgated by Petrus Ramus in the mid 1500s, aimed to expunge the rhetorical flourishes and (perhaps overly) intricate lines of theological thought common to the Scholastic writings that had long maintained dominance in the European universities. In order to do so, Ramus argued that his logically system had been embedded in our world through the mandate of God. By continually asserting God as a universally beneficial teleology (that is, there is no chance of him acting like Descartes’ deceptive sense-data demon) Ramus was able to assert “that knowledge consists of sets of mental items, thereby [implying] one-to-one correspondence between terms and things,” dispensing with much of invention and rhetoric and leaving dialectic naturalized in their wakes. ³

The assurance of a discourse community that shares the Ramist conception of language acts as a sort of transcendental guarantee that certain tropes, structures and forms will be understood without over-explication on the part of the author, leading to a unique type of meaning-stabilizing context in Puritan thought. Because, according to Ramus, “The ende, of naturall thinges is man, and of man God” language needs no calisthenics in order to properly represent the world.⁴. All things, including the ratiocinations of man, are ruled by the law of God, and simply by reasoning correctly one will communicate effectively. As Miller notes, this phenomenon transfers even into the more literary writings of the Puritan elite:

“To a Ramist rhetorician, verse was simply a heightened form of eloquence, it was speech more plenteously ornamented with tropes and figures than prose, but still like speech; like the oration, its function was to carry inartificial arguments from man to man.” ⁵

Correspondence with God’s world was the chief end of language, and Ramism was the royal-road means by which this goal was achieved.

However, the assertion of such a shared practice does not necessarily make it so. The question of how to accurately spot examples of some sort of unified plain style remains open. There perhaps may not even be a consistent deviation from prose standards of the time to be traced in Puritan, Ramism-inspired writings. By delineating standard parameters for texts of this period and then systematically searching for where deviations occur methods like power law inference can supplement historical investigation, offering insight the unaided mind can often grasp (“oh, this text seems like plain style”) by investigating the factors that might trigger such a phenomenal event in the first place (in this case, the presence of a value of a that differs from the norm). To that end, I used zipfexplorer to produce studies of the works of two individual Puritan thinkers (Thomas Hooker and Cotton Mather) as well as an additional set of results consisting of miscellaneous texts that can be classified as at least vaguely Puritan. In addition, I produced a study of works by non-Puritan or not manifestly Puritan writers of the same basic era. Texts were collected from the Chadwyck-Healey databases, as well as Project Gutenberg. I then plotted the value of that resulted from fitting these texts to a power distribution to a scatter plot with the Y axis representing value and the X the number of tokens (roughly, words) in the document used. ⁶

In this first plot of the texts initially assumed to not represent the plain style, we can see that most of the texts do cluster around the ‘normal’, per Cancho, of about two. Some outliers exist (and indeed, Defoe’s style was plain enough to be adored by Protestants henceforth) but a value above 1.9 seems to be average across this era. Compare these plots to those of our representative Puritans:

In this case, each one of these writers has topped out at a value of ~1.9. Rowlandson and Bunyan dip almost to 1.7 mark reported by Cancho to be characteristic of Military texts. Thomas Hooker finds himself writing in a similar frame of mind:

Hooker, a writer whose use of Ramism’s bracket-like disjunctive logical structure could be understated as “extremely liberal” never reaches the 1.9 mark. On the other hand, the works of Cotton Mather remain mostly clustered nearer the “normal” 2 mark and never dip into the 1.7 range:

Out of our computational investigations, both a pattern and an exception emerge. On the whole, the texts considered representative of the plain style produce significantly lower values than the period texts with the “not plain style” null hypothesis. ⁷  Modifying Cancho’s claim, we might say that these Puritan writers aren’t undervaluing the utility of their communication — rather they are relying on a philosophical structure (Ramism) as well as the knowledge that those who participate in their literary community share belief in this philosophical structure in order to offload much of their communicative work to a shared semantic umwelt. With Ramist dialectic in hand and mind Ramist writers can avoid over-elaboration in their writing without fear. The transcendental signifier of God will assure the uncorrupted transmission of their meaning, and their certainty that their interlocutors share the same intellectual context will assure that these readers possess the right methods for discovering this incorruptible meaning. This historical regime of meaning then loses its specific phenomenal force when read by modern interpreters. Whereas the “sense” or phenomenal experience of a low value of a may have once transmitted the conditions of its own interpretation to its readers (“treat me as a Ramist text!”) in the modern era that lacks not only belief in God as the unifier of word and object but even knowledge of Ramism itself this phenomenal experience might find other explanations (“this is an archaic text”). By correlating the power law data with historical knowledge we at least glimpse the reception of these texts in their time (and can, perhaps, more accurately track modern texts influenced by their strictures).

Strangely, Cotton Mather, arch-puritan and defender of the Salem witch trials does buck the low a trend to some degree. Though this initially seems a blemish on the hypothesis promulgated above, Mather’s own historical idiosyncrasies account for his uniqueness in a way that makes the exception prove the rule. While other Puritan intellectuals mostly shunned the developments of natural science that trickled over from Britain and continental Europe, Mather heartily embraced them, corresponding with the Royal Society, writing books about science and medicine (The Christian Philosopher and The Angel of Bethesda respectively) and even defending the use of inoculation during a smallpox epidemic in 1721. These facts would make it far too easy to laud both Mather and his prose as representatives of a coming American modernity where lucid, exploratory writing takes the place of reliance on theological obscurantism. Rather than follow that tack, I want to draw on Robert Tindol’s assertion that we should “move beyond the outdated opposition between Enlightenment and religious thought” in the work of Mather and instead focus on how Mather “was able to reconcile his religious views and scientific adventurousness in a way that diminished neither.”⁸. While Tindol goes on to close read the spots in Mather’s corpus where theology and natural science blend, discovering that Mather in fact tends prioritize scientific results over long-held theological beliefs, triangulating our results through the history of Ramism provides a slightly different conclusion. Due to a lack of fair, typed in copies of any of his specifically scientific writings, my Mather sample contains only religiously-attuned works. Yet even though none of these works are included in Tindol’s survey, they all register a higher a than the other Puritan texts. Mather’s stylistic “modernity” extends beyond just The Christian Scientist and Angel of Bethesda. Indeed, looking at the problem from this perspective reveals a shift that entails more than rectifying science and theology. Perhaps through his encounters with natural science (and other facets of European modernity) Mather develops a style that sees problems in general (crossing the fields of pathology, natural science, and yes, even theology) as more local and independent. This means that they require some extensive teasing out within the bounds of Mather’s texts proper in order to communicate them effectively. Natural science’s focus on local conditions and histories rather than overarching frameworks almost necessarily requires this type of communicative specificity. This comes as a definitive shift from earlier Ramist thinkers who could rely on the overarching telos of a meaning-assuring God, as well as a knowledge of Ramism-infused method, to make a point without explanatory acrobatics. In this way Mather shifts away from Cancho’s ~1.7 mark and towards the multiplicity of intellectual frameworks associated with modern cosmopolitanism and transnational cultural exchange.

While this post has focused mainly on texts below the “normal” a point of 2.0, the next, looking at experimental writing from Stein to the modern era, will tarry with those that go above.  This will help to show that this technique is non-vacuous: there are significant variations in a between texts that might have otherwise been lumped together as “experimental,” offering a way to track influence and shared semantic realms that phenomenal experience of a reading might not straightforwardly reveal.

Text Listings:

(Name, token count, a)

Assumed Not Plain Style

Assumed Plain Style

Hooker

Mather