Micro Reading Galileo’s Assayer, Ariosto’s Orlando furioso, and Tasso’s Jerusalem Delivered
How much does the natural philosopher owe to his favorite Renaissance epic poet, Ludovico Ariosto? Galileo’s affinity for Ariosto is well documented, including his patient edits to Ariosto’s epic poem Orlando furioso. Computational text analysis can move beyond Italo Calvino’s aesthetic reader-response evaluations of Galileo’s Ariostan prose and answer the question: just what does his writing share with Ariosto’s poem (published nearly a century prior to Galileo’s work) or with Tasso’s poem (published and debated while Galileo was a young man)? But computation does more than demonstrate just those commonalities. Exploring even something as basic as two-word (bigram) or three-word (trigram) units in the texts reveals what they do not share, and facilitates exploration of how Galileo’s writing related to other popular reading materials of the period.
The examples that follow are the results from drafting a method for examining this question.
What does a treatise on comets sound like?
Galileo’s Saggiatore/Assayer (Rome, 1623) is a delightful, literary-inspired criticism of a contemporary opponent’s views on the nature of comets. Derivations of the word comet appear once every 300 words, relatively often. If we think of this as a regular rhythm in the text, it would look like the diagram below, which counts a comet “beat” at every 300th word in the Assayer:
Although that average gives us a sense of how much content is devoted to comets in the text, its actual appearance is much less regular, to the point of even disappearing for large sections of the text.
What can a method like this reveal about Galileo’s linguistic similarity to Ariosto’s epic poem or Tasso’s chivalric poem, two of the most hotly contested examples of literary expression during his lifetime? What are Ariosto’s words? What are Tasso’s? Or, which words are in one poem, but not in the other? Computation offers a fast way to differentiate the lists of words that occur in each poem (similarly to how Excel can quickly detect duplicates in a column). The results reveal that 8% of the word usage in the Gerusalemme liberata is distinct from the Furioso, while 13% of the Furioso‘s word usage is distinct from Tasso’s poem (these are expressed as percentages to account for the different lengths of the poems). For computational text analysts, the details of my method are: make all text lower case, remove punctuation, create a vector of words for each poem. The results are in the table below (“Unique” refers to tokens and types that occur in one poem but not the other):
These unique words range from lexical families (Ariosto uses various forms of “abbarbagliar” (to dazzle), but Tasso does not use the verb at all) to “stop” words (articles, pronouns, prepositions, conjunctions, and negations). For example, Ariosto uses the articulated preposition “agl'” and “agli”, but Tasso does not; Tasso uses “nella” and “nelle”, but Ariosto does not. This comparison alone is a fascinating result that I have to gloss over now. As the co-authors of the first Stanford Literary Lab Pamphlet summarize in their work on “Quantitative Formalism” and the linguistic markers of genre, “There is something paradoxical in these traits that classify so well, and explain so little.”1
Given that Galileo was so critical of Tasso, my preliminary hypothesis was that the words unique to Tasso did not occur often in Galileo’s treatise on comets.
As it turns out, 1 in 42 words in Galileo’s treatise is a word present in Tasso’s poem (but not in Ariosto’s). Given Galileo’s intellectual development alongside the publication of Tasso’s work, this frequency of both content-bearing vocabulary and stop words seems almost to be expected. This is the language more contemporary to Galileo. So would the connection to Ariosto be stronger or weaker?
Even after removing false positives such as “mario” and “grassi”, both of which are names in the treatise on comets, uniquely Ariosto words occur with a frequency of 1 in 13 in the Assayer (1310 words that occur a total of 5,162 times). This is particularly important in the case of the “abbarbagliar” family of words (loosely translated as ‘to dazzle’ but with an etymology that reveals the earliest theories of optics in which images emanated from an object and penetrated the eye), a verb that Ariosto reserves for the knight Ruggiero’s magic shield that shines with a light so bright it incapacitates all who see it. Importantly this magic shield is only used on monsters in the poem, and it is considered unchivalrous to use it against other knights. In fact, when Ruggiero accidentally uses it on his companions he casts the shield into a lake so that it can never disgrace him again. In the Assayer, having this intertextual context, gives new meaning to Galileo’s defense of his student Mario Guiducci to his opponent (writing under the name of Lothario Sarsi): “If Sarsi will permit me to be frank, nothing more need be said here than that he has tried to dress up Sig. Mario’s remark so as to dazzle [abbarbagliar] the reader’s vision and leave him with the idea that Sig. Mario had made a blunder.”2 I would argue that Galileo is absolutely using the term with this chivalric connotation in his suggestion that Sarsi (and therefore Grassi) is trying to rhetorically stun his readers so that they cannot see the erroneous philosophical logic that lies behind his words. That leaves 1,309 similarities left to explore.
This preliminary analysis raises more questions than it answers. I am not sure to what extent the poems of Ariosto and Tasso can stand in for the linguistic features of literature of their decades. How much variation existed in the use of stop words (particularly the articulated prepositions) between editions of each poem? That is, to what degree can the stop words be trusted as unique linguistic features in the early modern period?
As far as the relationships between Ariosto, Tasso, and Galileo, the next steps are to examine what a study of hapaxes, bigrams, trigrams, and fourgrams might reveal.
 Sarah Allison, Ryan Heuser, Matthew Jockers, Franco Moretti, and Michael Witmore, “Quantitative Formalism: An Experiment” Stanford Literary Lab Pamphlet 1 (Jan. 15, 2011): 24. Burrows, Computation into Criticism: A Study of Jane Austen’s Novels and an Experiment in Method (1988) argues that idiolect is predicated on stop words, not those with “distinctive semantic content.”
 Galileo Galilei, Assayer, trans. Stillman Drake, in Controversy on the Comets of 1618, ed. and trans. Stillman Drake and C.D. O’Malley (Philadelphia: University of Pennsylvania Press, 1960), 187.