Acts of Aesthetics: Publishing as Recursive Agency in the Long Eighteenth Century
Of primary concern to late eighteenth-century society was the sheer volume of printed work being produced in England. The response of authors troubled by this perceived crisis was an outpouring of works on taste, aesthetics, genre and literature, which attempted to describe and provide corrective solutions to the problem of over-publication. Yet this response itself, of course, only added to the number of works within Britain. How did the solution to a deluge of print become more printed materials? Did these authors envision print’s agency as a self-corrective process? And, if so, how can we recover and perhaps even model the connections between works that encompassed the responses to print? This essay outlines a potential solution to this problem by sampling a highly focused selection of digitized texts that raise the issue of over-publication. Through a quantitative analysis of these texts, it identifies the lexical patterns that may reveal the ways in which authors of the period envisioned the work of aesthetics. In particular, this article identifies an emerging consensus on how printed objects became agents within the socio-cultural world of the long eighteenth century: both as objects which could act on readers and as objects which could act on other printed texts. By comparing the language that these clusters of texts deploy to discuss the agency of print to their traditional generic, theoretical or historical groupings, we can begin to examine the process by which the potential power of print became the solution to the dangers it, itself, presented.
The following is an experiment in understanding the relationship between the practice of aesthetics in the eighteenth and early nineteenth centuries and the transformation of the technology of print. Contemporary theories of Enlightenment and post-Enlightenment thought most often focus on the philosophic or cultural goals of aesthetics, while media-oriented studies use the rapid increase and increasing availability of print to read coextensive changes into eighteenth-century economic, political and social practice. Yet, aesthetics, as a system-theory of artistic production, is clearly bound up with the material realities of such production and is therefore shaped, to a great extent, by the popular and intellectual responses to changes in the means and mode of how the objects it purports to theorize are produced. Consequently, efforts to understand aesthetics as a disinterested philosophic endeavor (or even as a discipline that responds to social and cultural, but not technological concerns) are unable to fully reconstruct the development and transformations of aesthetic philosophy during the Enlightenment. Ernst Cassirer articulated this problem early in the twentieth century in his Philosophy of Enlightenment, claiming that Enlightenment aesthetics should not be understood simply as a subspecies of organizational philosophy but instead as a search for the unity between what he calls “demand and act” or “artistic form and reflective contemplation” (Cassier 278). According to Cassirer, “one of the imperishable titles to distinction of the epoch of the Enlightenment” is that “it joined, to a degree scarcely ever achieved before, the critical with the productive function and converted the one directly into the other” (278). If we are to take his claim seriously, it becomes necessary to unite our ideas on reflective contemplation (aesthetics) and the single greatest media-transformation in the period: the rise of print. In other words, how do theories of aesthetics and, particularly taste, form a response to the popular awareness of the radically increased output of print during the late eighteenth and early nineteenth centuries?
To suggest that the response to print was mixed, is to greatly understate the central problem of this question. Much like the field of the responses to our own moment of media transformation, the responses to print in the eighteenth century comprises a complex network of social, technological and ethical concerns that resists a reduction to a simple positive/negative dichotomy. Take for example Freidrich Karl von Moser’s essay on Publizität (literally “Publicity” but here referring to the explosion of printed materials in Germany in the late eighteenth century), published in 1792.
The torrent of publicity, in the good and bad senses, can no longer be stopped. It has been allowed to go too far. It should have been dammed up long ago and diverted onto another course. No one took the embers seriously because they were covered with ashes. The inner fire was disregarded because no one saw flames, or everyone thought they could be extinguished easily. All the lamentations, all the settlements of terms and committee resolutions with their demands, promises, and threats, come much, much too late. Given the entire constitution of the disharmonious system of the empire, given the laziness, selfishness, and powerlessness of so many greater and lesser estates, each so different in respect to abilities and intentions, given the whole character, politics, and independence of the book trade, the liberty and insolence of so many writers, and the insatiable lust to read of all estates, such lamentations and demands do exactly as much good as the well-known proposal of General von Kyau: that one should pave the meadows, so that moles could not harm them.
The times have passed and it is too late to try to shut out the light. The longer it goes on, the more it comes to this: whether this light should only illuminate and enlighten or ignite and inflame?
Here, Moser compares the increase in print to an uncontrollable fire that is at risk of consuming society: his clear call that the “torrent of publicity, in the good and bad senses […] should have been dammed up long ago and diverted onto another course” seems to argue for a cessation to publication wholesale. Yet, his response enters into nuance when he addresses the question of: “whether this light should only illuminate and enlighten or ignite and inflame?” raising at least the possibility that certain works contain the promise of being socially redemptive (Moser 115). Similar conflicting understandings of print technology were also operative in England during the same period: compare Constantina Grierson’s 1764 poem “The Art of Printing” with William Ireland’s “Of Foolish Unprofitable Books” from his Stulfitera Navis. On the one hand, Grierson celebrates the geographical and chronological range of printed materials (which become a technological extension of the author, able to traverse both time and space). In her poem, books contain both the solution to immortality and resurrection as she identifies them as “dead letters thus with Living Notions fraught” (Grierson 11). On the other hand, Ireland’s poem repudiates the immorality of printed materials in a familiar theme that links increased production and availability with increased sensationalism and immodesty. Yet even Ireland limits his argument to “unprofitable books,” forging an inverse relationship between economic and social profit: the worse a book is for the Ladies of Britain, he argues through the poem, the more likely it is to make its author famous (Ireland 1-24). Again, Ireland couches his argument in carefully nuanced terms: he recognizes the impracticality and danger of curtailing the freedom of the press in his opening lines and instead presents a case against particular popular genres.
Finally, the complexity of response to printing technology operates intra-generically, revealing in full the space that separates Cassirer’s contemplation from action. Compare Charles Abbott’s Essay on the Use and Abuse of Satire to the anonymously published Folly, A Satire on the Times. For Abbott, like many thinkers ranging from Milton to Kant, the press secures the public from the abuses of government: specifically the critical enterprise of satiric writing, when combined with the egalitarian dissemination supposedly inherent in printing technology, allows the public to participate in the political system. In his words, “By the invention of the Art of Printing, the dissemination of Political Satire has been eminently facilitated, and in our own country it possesses an importance as little known to the surrounding nations, as the peculiar privilege by which its freedom is secured” (Abbott 5). Note also, the national focus in his argument: it is Britain’s unique relationship to the press that has created its exceptional public liberties. Abbot here is expressing a proto-Habermasian belief in the processes by “which the state-governed public sphere was appropriated by the public of private people making use of their reason […] functionally converting the public sphere in the world of letters already equipped with institutions of the public and with forms for discussion” (Habermas 51). Conversely, the anonymous author of Folly uses the very genre highlighted by Abbott to make the inverse point about the technology: “While classic bards are styl’d unwise / And scribblers hail’d for scribbling lyes, […]On town and country let us live, / By DUPING we can only thrive” (Folly 7-8;15-16). Here, the institution of printing is that which limits the freedom of the public sphere by indulging the taste for entertainment even while it disseminates confusion and untruth.
The possibility of finding our way out of the circular logic that underlies the movement between these two texts is what I argue is at stake in this investigation. Abbott’s text makes the claim that print technology can be used to enable satire’s critique of governmental, social or cultural institutions, a critique that the author of Folly employs to condemn the very technology that underwrites his ability to condemn it. The synthesis that unites the texts which criticize print in the eighteenth century is their dependence upon the very medium they seek to limit: to reformulate Cassirer’s theory of aesthetics with which we began, we can identify a subject/object unity in the negative responses to printing technology that calls into question the use of the very medium under critique to stage the critique. How did Enlightenment authors understand this vicious cycle of printed critiques of print to aid in solving the problem of the overload of print?
This question becomes even more loaded, when we turn from the total field of responses to print technology to the area of knowledge whose very purpose was to frame a response to human media production: the genre of aesthetic theory. The close relationship, indeed the interdependency, between aesthetic theory and literary creation is frequently glossed over in favor of a theoretical investigation of philosophic principles (such as has been offered by post-structural critiques of aesthetics) or in pursuit of the ways taste functions as a method of social control in Marxist or cultural reading (as offered, for example, by Bourdieu (30)). Nevertheless, even a cursory examination reveals the deep connections between writing as an object of study and aesthetic theory as the genre that studies it. Burke’s Enquiry pulls its examples of the affective sublime, not from nature, but from the work of Milton, Dryden and Shakespeare, before concluding with an interrogation of the special functions of words in aesthetic representation (Burke 128). Likewise, Gilpin’s tour of the River Wye in his Observations Relative Chiefly to Picturesque Beauty, employs Paradise Lost as a touchstone for aesthetic experience (Gilpin 106). Even German aesthetic tradition holds writing to be a special case of human production: witness Lessing’s theory of ekphrasis in the Laocoön which seeks to textually explicate the visual arts or Hegel’s Aesthetics which places writing (specifically philosophic writing) as the limit and endpoint of art. If aesthetics and theories of taste are an attempt to organize, categorize and limit the range of responses to art, then their focus on writing suggests the possibility of an underlying cognizance of the crisis of print. This follows from the expansion of the technology itself: the need for a theory of taste to control the range of responses to aesthetic objects is a response to the crisis occasioned by the exponentially increasing availability of just these objects, mass produced on an unprecedented scale. Yet, if in fact these texts respond to this problem, they again do so by harnessing the medium itself as a technology that can limit its own overproduction. The ability of print to widely distribute the opinions, theories and philosophies of the eighteenth-century aesthetic theorists (as acknowledged as early as the tenth issue of the Spectator when Addison accounts for his probable readership (36)) underlies any attempt of their part to critique the medium itself. Again, we are left with the question: how does printing more works offer a solution to the problem of too many printed works? Read another way, what can print DO, what action can it take self-reflexively, such that it can form its own limitations through multiplication?
One of the central difficulties with questions such as this is that our potential for response is already limited by the conditions of the historical situation itself. The exponential increase in the number of printed texts during the latter half of the eighteenth century has become a familiar narrative to literary studies. Generated from the data contained within the Eighteenth Century Collections Online (or ECCO) database, this chart, Figure 1, which can be generated from any representative sample of printed work during this period, is likewise a familiar visual. The attempt to investigate the printed response to print, even just in works that can be considered theories of taste or aesthetics, across the span of the eighteenth century and Romantic period runs into the same situation as faced by the writers of the period themselves: there are simply too many works to adequately comprehend or represent through our traditional strategies of close reading. Moreover, to adequately recover if and how theories of taste and criticism responded to the transformation of print technology, we must enact a forgetting of our received history of aesthetics. Whether under the rubric of philosophic theory or social/cultural studies, a certain unity is presumed among the aesthetic thinkers of the eighteenth century, such that while we can trace a genealogical progression, each thinker is responding to the same set of concerns plus those of his predecessor. Even Cassirer, as he reformulates the goal of eighteenth-century aesthetics as a synthesis of thinking and doing, relegates aesthetic writing in England to a set of footnotes and responses to Shaftesbury, who died in 1713 (Cassier 313). If aesthetic theory did respond to the increase in printed materials during the eighteenth century, particularly in Britain, then we should witness a shift in writing whose existence can be traced back to the increase in print itself, separate from and in addition to any critical transformations offered by any individual author.
Texts Published in the Eighteenth Century (from ECCO)
What I want to offer here, however, is not a complete solution to the questions of aesthetics, taste and print that I have raised. Instead, I want to reframe the question such that we can approach the possibility of reconstructing the effects of the increase of print on aesthetic writing in a way that potentially navigates the two difficulties described above: the sheer volume of printed work and, more problematically, our implicit biases as literary scholars towards narrative histories of aesthetic theory and philosophy. A turn to the new computational tools available to literary studies, when combined with the traditional forms of literary analysis, could, I argue, provide a potential methodology through which these problems could be addressed. Not only do statistical analyses of lexical data mined from digital archives allow scholars to parse large samples of texts (particularly within eighteenth-century studies given the extensive digital archives of period texts), but these types of analyses can reveal correlations between texts and patterns among works that the critical reading strategies of the humanities are ill suited to recognize. Rather than reading for the substance of each work, these methods can allow us to examine the linguistic traces that certain texts on aesthetics share: this, in turn, gives us access to the way in which these authors conceived of aesthetics, rather than their specific aesthetic theories themselves. By looking at patterns of language usage across the historical spectrum of the Enlightenment as they are contextualized within large samples of generically or thematically similar texts on aesthetics and within individual works themselves, we can begin, I argue, to reconstruct the burden of socio-cultural work that eighteenth-century authors placed upon published texts on aesthetic theory. This method allows me to uncover the ways in which particular authors and groups of authors responded to the intellectual and social pressures inherent within the ideas of writing, publishing and taste. What I want to offer in this paper, using this combined quantitative and critical methodology, are two potential critical approaches to the contradictory use of print media by eighteenth-century authors to help stem the spread of print.
To uncover these underlying patterns within eighteenth and early nineteenth-century aesthetic texts using computational analysis, a set of metrics is required: these can be derived through new strategies of text mining when performed on a carefully selected sample from works from the period under investigation. “Text mining” encompasses a set of quantitative and statistical tools that function by locating metrics based on numerical patterns within the language and syntax of texts. By extrapolating these patterns, and then applying a set of statistical operations and transformations to the resulting metrics, a text mining package is able to locate resemblances between texts that are lexically or syntactically robust, but not necessarily apparent thematically or critically. For the purposes of this study, I will be using the “tm” (or text mining) package in the open source R software environment for statistical computing (GNU 2010). In order to keep the results of this preliminary investigation manageable, I will limit the data under investigation to 28 digitized works of eighteenth and early nineteenth-century aesthetics and theory: because of the exploratory nature of this investigation, a large sample size could introduce excess noise that could obscure the appearance of potential clusters. Other areas of literary studies that have pioneered similar quantitative methodologies have found that a careful analysis of a limited, critically selected sample, in studies of stylometrics or authorship attribution, for example, can be most successful in clustering like texts (Hoover 180, 200-201; Jockers 217-218; Yu 328). By quantitatively clustering a critically selected sample of texts on aesthetic theory, my goal is to uncover the potential for patterns of language usage that point toward shifts in the language of aesthetics that correspond to the historical phenomena of the expansion of print. While this sample only represents a small subset of the large number of works written on aesthetics available in digital formats, the exploratory nature of this pilot study assumes that the need for clarity surpasses the potential for comprehension.
The texts of this sample were selected from a number of online databases available in our field, including ECCO, LION, Project Gutenberg and Google Books. In order to ensure that the sample was comprised of works on aesthetic theory that were representative of the period and yet still evidenced some cognizance of print culture, I used a set of hierarchically weighted terms to identify key texts. Texts in which these terms appeared most frequently, as well as texts which used these terms in their title were selected as representative examples of the theory under investigation. To simplify this initial study, the sample was limited to works published in English. The resulting subset of texts represent a core sample of aesthetic writing throughout the eighteenth and early nineteenth centuries: the works range from 1701 to 1829 and cover an range of aesthetic and critical writing. Using the “tm” package’s “DocumentTermMatrix” function, I was able, using this sample, to generate a table that assigns all unique tokens (or words) in each text a metric based on their overall frequency, both within the sample and within the individual texts. With this set of metrics, we can examine how the language of each individual text affects its relationship to the other texts in the corpus.
To reveal the underlying patterns within the Document Term Matrix, we can begin with an exploratory model of the data using a hierarchical clustering of the raw frequencies plotted in a dendrogram (Figure 2). This dendrogram is generated by the R package Hclust and shows each text grouped according to its hierarchical relationship to all of the other texts in the sample. The closer the relationship between two texts, the lower on the y axis the two texts are connected. The higher the connection, therefore, the less the lexical elements of the text overlap in their arrangement. The dendrogram produced from my initial data on these texts divides into three groups. The group on the far left is composed of outliers: the texts in this population were initially all an order of magnitude longer than the other texts in the sample. Although some attempt was made to regularize the length of the texts, the regularization clearly affected the composition of these works such that they are unrelatable to the other members of the sample.
The remaining texts fall into two clear groups. While there is no obvious thematic or generic regularity that would indicate the principle of selection, the metadata on these samples reveals an interesting pattern. While there is some overlap, allowing for some linguistic variability between the two groups, on average, the texts in middle group were composed twenty years earlier than the texts on the right group such that the average date of composition of the first is 1757 and the second is 1775. These results suggest that the language of aesthetic composition experiences a substantive change within the eighteenth century and, moreover, give us a chronological outline through which we can begin to pinpoint these transformations. This observation leads to the first potential conclusion that I want to draw from this data.
This time frame for the transformation of the language of aesthetic theory provides us with a key to understanding this evolution as an effect of print technology. When we map the chronological limits of this change, such as they exist within the model, back onto our map of the number of works published during the eighteenth century (Figure 3), these chronological limits coincide with the beginnings of the radical increase in print during the 1760s and 1770s. What we have identified using this quantitative model, therefore, is a fuzzy boundary between two groups of texts: those written before the print explosion of the later eighteenth century, and those written after. This is not to suggest, of course, that this model refuses the possibility that authors writing before this boundary were not aware of or concerned with the “problem” of print: clearly that is not the case as many of the works that I used to outline the problem take print as their subject and were composed prior to 1760. The model does not correlate precisely with history, nor is it possible to create a model that does so. Rather, this transformation seems to indicate a change in the way aesthetic authors wrote about print in the later and post-Enlightenment, or, at the very least, a change in the intensity of their concern for print as a socio-cultural phenomenon. If we were to retrace the historical contingency of this period of transformation, we can, of course, locate a key event in print history that falls within the limit of this transformation. In 1774, the case of Donaldson vs. Beckett, over the right of Scottish printers to print runs of James Thomson’s The Seasons, repealed the common practice of perpetual copyright and reinstated the 14 year limit on copyright set by the Statute of Anne in 1710. In his socio-economic history of book printing, selling and reading, William St. Clair identifies this as a transformative moment within the reading public of the eighteenth century: this is the point at which inexpensive, accessible books began to flood the English marketplace creating what he calls, the “brief copyright window” before social and political forces responded to this freedom with a new set of political controls designed to discourage and contain what William Pitt the Younger referred to as “cheap literature” (The Reading Nation in the Romantic Period 486-7). This model, therefore, allows us to tentatively read the development of aesthetics back onto the socioeconomic concerns of printing, suggesting the formulation, in a different configuration, of what Clifford Siskin has referred to as the “political economy of taste” (115). Moreover, as the shift that St. Clair describes in the reading public of England involved the increased participation of the lower classes in the public sphere, the shift in theories of taste might indicate a reactionary response to the opening of the reading public to these new groups. The historically based clustering that we can interpret in this model could therefore be evidence of a class- based response to the change in print technology: this aligns the aesthetic theory under investigation here with Roy Porter’s hypothesis that taste “made the new social adhesive which would cement propertied élites together” (Porter 195). In all of these accounts, taste, as a sub-discipline of aesthetics, alters its mode of operation in response to a significant change in the material production of its subject. The model, unfortunately, does not provide a one-to-one correlative field onto which we can read this transformation precisely: what it does do is allow us to reframe the question “how does aesthetic theory change in response to print technology?”, instead allowing us to ask: “how does the copyright window affect the language of aesthetics and what does this tell us about the ways in which aesthetics covertly responds to changes in print production?”
Texts Published in the Eighteenth Century (from ECCO)
The quantitative nature of this model allows me to expand the tentative analysis beyond simply identifying the historical continuities between changes in printing and shifts in the language of aesthetics. We can also interrogate the model to potentially identify the nature of the shifts in language that took place in the later eighteenth century. As the texts are clustered based on the frequency of particular words in specific patterns, we can look at the criteria for clustering in order to determine what about the language of these texts caused their separation into discrete groups. What linguistic elements, in other words, changed in the ways in which authors wrote about aesthetics during this period? Similarly, since we are concerned with particular configurations of key terms (how authors of texts on aesthetic theory addressed questions of taste, print, authorship, reading and writing) we can further interrogate our model to see the effect of these particular ideas on the quantitative analysis. While a variable selection procedure would allow us to identify the most informative variables in the clustering procedure overall, the historical grouping of the texts within the dendrogram allows us a different way to access the principle of their categorization. As these texts are separated based on the absolute differences inherent in their language usage, terms that are frequent in one set, but absent in another would be among the most informative identifiers of group selection. By dividing the sample into the two groups indicated by the dendrogram and then individually identifying the 30 highest frequency terms in each sub-sample (Figure 4), we can locate potential signifiers that are responsible for the categorization we have witnessed.
Most Frequent Words (Group 1)
Most Frequent Words (Group 2)
As there is much overlap of language usage between the two groups of texts, as we would expect, many of the high frequency terms are duplicated. In order to explore the key differences in the two texts, I have highlighted the unique words: the terms that occur frequently in one group, but are significantly less frequent, or absent, in the other. Beyond their shared terms, there are a number of key differences between the data sets. In the first group, the unique words appear to be more descriptive in nature: words like “poetry,” “composition” and “art” represent a range of subjects for the theory, while “Poet,” “style,” “genius” and “imagination” provide a descriptive terminology for the process of creation. No similar groups of objects or actors appears in the second: instead, the unique terms seem to focus less on the creation of the aesthetic object by an agent and instead, indicate an interest in the “effect” of the aesthetic object on the “passions” or the “human.” As these terms suggest that the second group of texts is more concerned with the effect of objects (rather than their creation or evaluation), the appearance of the terms “act” and “action” within the second group take on special relevance: if the concern of the authors of this second group is for the ways in which the theory needs to account for the effects of aesthetics objects within a socio-cultural environment, then the action of the objects themselves is what is at issue. Moreover, the most frequent significant word within this group is “words,” suggesting that the objects primarily under consideration are textual artifacts whose distribution in print was a key concern for authors in the eighteenth-century. The increased frequency of “act” and “action” suggest, therefore that what is at stake in late eighteenth-century aesthetics is a species of action that could be related, in some way, to print. Their presence as unique terms within the second group potentially suggests, in other words, that in late and post-Enlightenment theories of taste, criticism and aesthetics, print comes to be thought of as possessing a kind of agency: both as something capable of acting on humans engaged in the act of reading, and as something capable of affecting other printed object. The assumption that the role of print could act as an agent within the long eighteenth century is commensurate with the identification of the agency of objects in social assemblages in Bruno LaTour’s Actor-network Theory and begins to point towards an explanation of how printed objects themselves could act as agents in shaping the interaction between reader and text and between texts. In a social assemblage, “the continuity of any course of action will rarely consist of human-to-human connections […] or of object-object connections, but will probably zigzag from one to the other” (LaTour 75).
Before using these results to aid in a critical reading of the texts themselves, it is worthwhile to test this assumption: if action does assume a new importance within the later Enlightenment and Romantic periods specifically in relation to the discipline of aesthetics and theories of taste, then the texts in our sample should show some variation based upon the their discussions of action within the context of our other key terms. To test this, we can single out the frequency scores of both “act” and “action” as well as the other key terms from which we derived the sample of texts (“write,” “publish/print,” and “taste,”) and explore how well the texts in our sample cluster based on these particular variables. Although we have already witnessed the effects of these terms as grouping variables in the dendrogram above (figure 2), the potential differences in the results based on the clustering algorithm used requires some additional testing before the validity of these conclusions can be assumed. As I am using clustering analysis here for exploratory data analysis (rather than as a methodology for testing hypotheses), modeling the data through a different form of statistical analysis will both aid in confirming the model suggested by the dendrogram and will create a new visualization of the data through which new patterns, not shaped by our reading comprehension or inherited history of aesthetics, can emerge. From these results I can return to the texts themselves in order to interpret the lexical patterns revealed within an intelligible framework.
As our initial analysis has provided us with two historical groupings of texts, we can use these pre-defined clusters to run a principle component analysis (PCA) and create a model in two dimensional space of the way that the two groups of texts use this particular set of words. A scatter plot of the resulting first two principle components (Figure 5) reveals that there is, in fact, a great deal of variation in the discussions of taste, writing and action across the long eighteenth century. While texts from the first group (predominantly early eighteenth-century texts) cluster tightly towards the left side of the graph, the texts from the second group vary greatly across the x axis. From this model, we can interpret two key conclusions. Firstly, if we assume that our sample of texts is representative of aesthetic writing across the long eighteenth century and Romantic periods, the use of “act” or “action” in relation to discussions of taste, writing and printing did change between the early and late eighteenth century. Secondly, the ways in which these terms were used varied greatly after this transformation, as evidenced by the radical change in cluster size. Authors of aesthetic texts did change their emphasis on the ways in which taste or authorship could constitute an action, but they did so in unpredictable ways. To some extent, this increase in variation can be explained as a de-regularizing of aesthetic theory in the transition from eighteenth-century thought to Romantic period practice: as Enlightenment systems of aesthetics became the Romantic discussion of taste, the terms of the debate became far less regularized within a particular generic form. Nevertheless, the inclusion of “action” within the reduced set of variables suggests that the previous analysis holds: within the fuzzy boundary between 1757 and 1775, the discussion of aesthetics as an action came to form a significant part of the discussion of taste and aesthetics.
Scatterplot of the First Two Principle Components
This leads me to a second conclusion that can be based on the experimental results. Returning to the texts themselves, it becomes apparent that the presence of “act” and “action” within the word groups is not simply an artifact of the clustering process. Throughout the texts that form the field of the second, chronologically later, group, there is an undercurrent of action: a focus on the ways in which art, or objects of taste, can operate within experiential reality, thus bridging the gap between subjective response and empirical action. For example, Henry Holme, Lord Kames’ Elements of Criticism, argues that “perceptions and action have an intimate correspondence” and goes further to suggest that aesthetic order forms the framework of thought and action itself: art acts as a control that prevents us from being “hurried from thought to thought, and from action to action, entirely at the mercy of chance” (30). While Kames does not make the connection between action and print explicit within his text, his focus on literary production as the object of criticism, coupled with our new qualified understanding of criticism’s investment in the changing medium of print, points towards a potential link between the two. Similarly, Clara Reeve’s Progress of Romance also explores the ways in which literary consumption leads to real-world action: however, unlike Kames, Reeve is explicitly addressing the effect of novels, potentially the most problematic object of print technology, on society. In her description of the ways in which a written work affects the mind, and therefore the actions, of the reader, Reeve is also careful to explore both the positive and negative potentials inherent in her formula: when faced with a work in which the love of glory has become an “object of contempt and ridicule” the reader could potentially “devote themselves to mean or mercenary pursuits which debase and corrupt the mind.” Whether positive or negative, the written (or printed) word can act as a “stimulus to excite men to action, and such as is the motive, such will the action be” (102). Her ultimate vindication of, carefully controlled, printed material within the text reminds us that our model of aesthetic language cannot be read as a simple binary: the responses to print in the late eighteenth century were not uniformly positive or negative. Instead, what appears to change is the description of how art, and specifically printed writing, brings action into being.
The slight overlap within the model’s historical grouping suggests, of course, that early eighteenth-century aesthetic texts, like their later counterparts, were also concerned with discussions of action; however, it is worth comparing the function of action within the poetics of the first group to understand the radical nature of the transformation we can witness in the second. John Dennis, in his 1704 text The Grounds of Criticism in Poetry, describes action as a function of passion which can be produced through poetic experience:
The subordinate End of Poetry, which is to please, is attained by exciting Passion, because every one who is pleased is moved, and either desires, or rejoices, or admires, or hopes, or the like. As we are moved by Pleasure which is Happiness, to do every thing we do, we may find upon a little Reflection, That every Man is incited by some Passion or other, either to Action, or to Contemplation; and Passion is the result either of Action or of Contemplation, as long as either of them please, and the more either of them pleases, the more they are attended with Passion
Two key aspects of this passage serve to differentiate it from the discussions of action within Kames’ or Reeves’ text. First, the action that gives rise to passion resides within descriptive poetics rather than real world articulation. Elsewhere in the text, Dennis attributes the creation of just this passion to the action described by Homer (Dennis “Specimen” n.p.). This is reflected in the list of potential actions that Dennis cites: “desire,” “rejoice,” “admire,” and “hope are all emotive actions that remain firmly within the sphere of poetic appreciation. Secondly, the cyclical nature of the process through which passion becomes action and, in turn, incites passion belies the possibility of a linear translation between word and world: not only can passion be sublimated into contemplation instead of action, but the end product of any action that results from poetic passion is, once again, poetic passion. The cycle that Dennis describes is different from the auto-generative network of printed responses to print: instead, like the action he references, his text is engaged with a purely descriptive criticism. The book does not advocate its dissemination in print as a potential solution to a perceived problem; instead, The Grounds of Criticism in Poetry belongs to the very contemplative response to aesthetic imagery that Dennis himself describes.
The transformation of aesthetic theory between the early Enlightenment and the late eighteenth century and Romantic periods is evident when we compare Dennis’ text with George Campbell’s Philosophy of Rhetoric . In Campbell’s work, which, in its entirety, links writing to rhetoric and print to writing, Campbell sets up a system of aesthetic response whereby knowledge affects fancy, which alters the passions and which, in turn, creates action: “Knowledge, the object of the intellect, furnisheth materials for the fancy; the fancy culls, compounds, and, by her mimic art, disposes these materials so as to affect the passions; the passions are the natural spurs to volition or action, and so need only to be right directed” (28-29). This clear chain of linkages provides a map through which, at least according to Campbell, the end result of art is action. Unlike Dennis’ cyclical progression, the structure of Campbell’s chain of affective response is linear: pointing away from the text into the world. Moreover, Campbell’s final dictum in this passage references the control and direction of the passions, and therefore of art (particularly writing) itself. As in the examples from Reeve and Kames, there is a link between art and action which suggests that if indeed we can recognize a new focus on action within aesthetic writing of the later eighteenth century, that focus could potentially come in the service of altering, or, perhaps regulating, the affects of print as it acts in society.
Again, we must return to the question of the mediation of these works and their awareness of the medium in which they reside. As printed responses to print, they actually participate within the system of response/action that all three articulate. Although the recognition of philosophy or critical writing as itself an aesthetic medium is perhaps best left to post-structural speculation, it is clear nevertheless that Campbell’s method of “directing” the proper set of responses to literary production is the work, or the action, of his text itself. Similarly, Reeves’ text, written in the genre of a dialogue, also functions as a system through which the production and dissemination of texts can be regulated. Both authors depend upon the very principles of action that they advocate in their own aesthetic theory to perform the work of this theory itself. As this self-reflexive system might suggest, the most effective field of action for a printed work is within the sphere of print itself. In other words, it is possible that the active power of print is effective enough that the only methods through which it can be limited lie within the set of active functions that it makes possible. This is a potential way to reconstruct the investment of authors of aesthetic or critical texts in print, even as they seek to regulate or control it. One advantage of this argument, as it is described by the model based upon our experimental data, is that it allows us to re-imagine aesthetics as a set of controls: philosophic controls on the reception of art and its place in the organizational system of Enlightenment thought; social controls limiting behavior based upon a set of moral and intellectual criteria; and technological controls seeking to limit the actionable possibilities of print through a redirection of its agency back towards its own action in the world. Moreover, thinking of aesthetics in this way allows us to reformulate Cassirer’s unity, with which we began. Rather than a synthesis of action and thought between two areas of human endeavor, as Cassirer suggests, we can instead rethink aesthetics itself as a method of social action that uses the technology of print to create a unity in itself between thinking and doing: adding our ideas of print to aesthetics makes thinking AS doing possible.
To conclude, I want to underline the experimental nature of this project. Not only is the method that I propose here in need of further refinement, but the set of tools that it provides are also an extension of a media technology itself. In this, it offers us an interesting case in which the new medium of textual surplus is applied to rethink the logic that governed the old, and thus must be thought of as part of the very system it studies. Nevertheless, the link between aesthetics, taste, print and action is strongly born out within the text themselves, suggesting that the patterns uncovered through such quantitative analysis reveals a key transformation within eighteenth-century and Romantic period aesthetics. As the over-production of print became a problem for readers and writers in the eighteenth century, aesthetic theory transformed from something that is done to something that does, or rather, something that the author creates to perform work within the economy of literacy in late eighteenth-century England. The very quantitative nature of the problem (the increasing number of printed texts that were designed to respond to, limit and control other printed texts) allows us to find a quantitative solution to the question of how additional print solves the problem of too much print. And yet by repurposing these tools as a method for critical analysis, we run the risk of repeating the work of the late eighteenth-century aesthetic theorists – the possibility of a media response to a problem with mediation is still at issue within our own critical and intellectual practice today. Only when an analysis of the self-reflexive agency of print is brought forward to our own critical practices will we be able to understand more fully the action of media, or even mediation, within our own field of practice.
Appendix 1. Sample of Texts Analyzed
A Discourse on Ancient and Modern Learning
Essay on Original Genius
Essay on the Nature and Constitution of Epic Poetry
Essay on the Study of Literature
Characteristics of Men, Manners Opinions and Times
Clio or A Discourse on Taste
Conjectures on Original Composition
Course of Lectures on Oratory and Criticism
Elements of Criticism
Essay on Taste
Essay Upon Writing
Essays on the Nature and Principles of Taste
General Introduction to the Waverly Novels
Grounds of Criticism in Poetry
Lectures on Poetry
Lectures on Rhetoric and Belles Lettres
Lectures on the English Poets
Miscellaneous Observations on Taste, Genius, Good Sense
On Fable and Romance
On the Cultivation of Taste
On the Origin and Progress of Novel Writing
On the Standard of Taste
Preface Comprising a Short Dissertation on the Novel
Preface to the Words of Shakespeare
Progress of Romance
The Advancement and Reformation of Modern Poetry
The Philosophy of Rhetoric
Mark Algee-Hewitt is a Mellon Post-doctoral fellow at McGill University. His work examines the relationship between literature and aesthetic theory of England and Germany during the long Eighteenth-century and Romantic period, bringing new quantitative methods to bear on large digital archives of period texts. He is currently working on a book that applies this approach to the history of the sublime.
The former approach is exemplified by work on aesthetics categories, such as the sublime or the picturesque: see, in particular, chapter 3 of Alexander Regiers’ Fracture and Fragmentation in British Romanticism or Judith Broome’s Fictive Domains: Body, Landscape and Nostolgia 1717-1770. William St. Clair help lay the contemporary ground work for the latter approach in his 2005 The Reading Nation in the Romantic Period and his recent work, for example his chapter on “Publishing, Authorship and Reading,” continues to explore the expansion of print as the primary factor in eighteenth and early nineteenth-century social changes. Recent work in the field of book history has also offered a range of media-oriented interpretive strategies for understanding Enlightenment thought: two very different approaches are offered by Andrew Piper’s Dreaming in Books and Section 2 of April London’s Literary History Writing.
The rise of “criticism,” broadly conceived, as a response to the expansion of eighteenth-century publication (a process exemplified by texts such as Arnold’s Culture and Anarchy) has formed the basis of recent accounts of the history of literary studies and writing in general (see, for example, Siskin’s The Work of Writing (54-78)). Also interesting is the continued understanding of the primary function of critical work as the necessary response to an overload of books as expressed, for example, in Susan Fisher’s editorial in Canadian Literature, “So Many Books” (7-8).
The total sample of 28 texts included 1,547,762 words: while it is reasonable to surmise that a literary scholar could analyze this quantity of language, he or she could do so based only on comprehension of the subject or idiosyncrasies of individual texts, not on word usage or lexical patterns among groups of texts. The strength of the quantitative aspects to the model I propose here lies in its ability to identify patterns of word usage that are not apparent within the text.
The specific search terms were: “publish,” “print,” “write,” “author,” “poet,” “scribbl*,” “press,” “literature,” “books,” “read,” “pen,” and “taste.”
At least three words had to appear more than five times in the text or once in the title.
See the list of texts, along with their associated document numbers in Appendix 1.
Using the tm package’s built in standardization functions for case, punctuation, stopwords and lemmatization, I reduced each text to its constituent set of significant words (these functions, in other words, remove the stems from root words and eliminate the set most common words across the sample [words such as “and”, “the”, “is” etc..])
Because many of the tokens appeared in just one or two texts, and were therefore not useful in clustering the sample as a whole, the total number of terms was reduced to the most significant with the removeSparseTerms function using a sparsity of 0.4. This removes all terms with a greater than 40% probability of having zero occurrences in any given document.
The dendrogram in Figure 2 was created using Ward’s method, an agglomerative clustering algorithm, testing all possible clusters and selecting those that minimize the error sum of squares (ESS).
A future solution to this problem might be found in the new strategy of affinity propagation, or AP, clustering. This technique has been shown to be effective in isolating representative parts of a text, such that the overall length can be reduced without affecting the meaning of the work (Frey and Dueck).
The numbers accompanying the terms indicate the number of instances of each word within the overall sample. The terms, therefore, are ordered according to frequency.
Sometimes several algorithms are applicable, and a priori arguments may not suffice to narrow down the choice to a single method. In such a situation it is probably a good idea to run more than one program and to carefully analyze and compare the resulting classifications, making use of their graphical displays. The interpretation of these results must then be based on insight into the meaning of the original data, together with some experience with the algorithms used
Kaufman and Rousseeuw 37
This analysis was completed using the SAS statistical analysis software package. A Principle Component Analysis (PCA) reduces the number of variables (in the case, unique terms) into a smaller number of variables responsible for the greatest proportion of variation. In this analysis, the resulting first two principle components were graphed on a scatter plot. While the 28 texts within this analysis would normally constitute an unacceptably small number of observations for a PCA, the reduction of the variables (terms) to five key terms (“write,” “publish” “taste” and “author” as well as “action”) maintains the general rule of significance that the sample must contain five times as many observations as variables (obs=28; var=5). Of course, increasing the sample size in future analyses will provide better results.
Anonymous. Folly: A Satire on the Times; Written by a Fool and younger Brother to Tristram Shandy. London: 1763.
Bourdieu, Pierre. Distinction: A Social Critique of the Judgment of Taste. trans. Richard Nice. Cambridge: Harvard UP, 1984,
Broome, Judith. Fictive Domains: Body, Landscape and Nostalgia. 1717-1770. Lewisburg, PA: Bucknell UP, 2007.
Burke, Edmund. A Philosophical Enquiry into the Origin of Our Ideas of the Sublime and Beautiful. Ed. James T. Boulton. Notre Dame: U of Notre Dame P, 1968.
Cassirer, Ersnt. The Philosophy of Enlightenment. Trans. Fritz C.A. Koelln and James P. Pettegrove. Princeton: Princeton UP, 1951.
Feinerer, Ingo, Kurt Hornik and David Meyer. “Text Mining Infrastructure in R.” Journal of Statistical Software 25.5 (2008): 1-54.
Frey, Brendan and Delbert Dueck. “Clustering by Passing Messages Between Data Points.” Science 315(2007): 972-976.
Gilpin, William. Observations, relative chiefly to picturesque beauty, made in the year 1772.London: 1786.
Habermas, Jürgen. The Structural Transformation of the Public Sphere: An Inquiry into a Category of Bourgeois Society. trans. Thomas Burger. Cambridge: MIT P, 2001.
Hegel, G. W. F. Hegel’s Aesthetics: Lectures on Fine Art. 2 vols. Trans. and Ed. T. M. Knox. Oxford: Oxford UP, 1975.
Hoover, David L. “Corpus Stylistics, Stylometry, and the Styles of Henry James.” Style 41 (2007): 174-203.
Jockers, Matthew L. and Daniela M. Witten. “A comparative study of machine learning methods for authorship attribution.” Literary and Linguistic Computing 25(2010): 215-223.
Kaufman, Leonard and Peter J. Rousseeuw. Finding Groups in Data: An Introduction to Cluster Analysis. Hoboken, NJ: Wiley, 2005.
LaTour, Bruno. Reassembling the Social: An Introduction to Actor-Network Theory. Oxford: Oxford UP, 2005.
Lessing, Gotthold Ephraim. Laocoön: An Essay on the Limits of Painting and Poetry. Trans. Edward Allen McCormick. New York: Bobbs-Merrill, 1962.
Von Moser, Friedrich Karl. “Publicity.” What is Enlightenment? Eighteenth-century Answers and Twentieth-century Question. Ed. James Schmidt. Berkeley: U of California P, 1996. 114-119.
Piper, Andrew. Dreaming in Books: the Making of the Bibliographic Imagination in the Romantic Age. Chicago: U of Chicago P, 2009.
Porter, Roy. The Creation of the Modern World: The Untold Story of the British Enlightenment. New York: Norton, 2000.
Regier, Alexander. Fracture and Fragmentation in British Romanticism. Cambridge: Cambridge UP, 2010.
Reeve, Clara. The Progress of Romance, through Times, Countries, and Manners, in a Course of Evening Conversations. Colchester: 1785.
Siskin, Clifford. The Work of Writing: Literature and Social Change in Britain 1700 – 1830. Baltimore: The Johns Hopkins UP, 1998.
St. Clair, William. “Publishing, Authorship and Reading.” The Cambridge Guide to Fiction in the Romantic Period. Ed. Richard Maxwell. Cambridge: Cambridge UP, 2007. 23-46.