In his keynote speech From Punched Cards to Treebanks: 60 Years of Computational Linguistics at the Eighth International Workshop on Treebanks and Linguistic Theories (Milan, 2009), Busa sketched three main typologies of informatics currently in use: “documentaristic” informatics, comprising all the informatic services allowing efficient “information retrieval”; “editorial” informatics, referring to the wide range of “multimedia” devices for reading books, watching films, browsing the Internet; and “hermeneutic informatics”, considering the “computerized text analysis, or language hermeneutics, i.e. interpretation, […] of all our ways of questioning the whys of language”  [Busa 1999, 5]. This latter was the informatic typology to which Busa devoted his inexhaustible attention during all his life. The focus on language, according to him, is strictly required by the nature of the relationships between man and computers, which interact by means of specific programming codes. Moreover, “the computer allows and exigently demands, as its specific capacities, an exhaustive, detailed, deep, quantitative knowledge, derived from huge amounts of natural texts”  [Busa 1999, 6]. As a consequence, it is arguable that computers and human beings, technological devices and humanities are not competitors or antagonists, but they are potential allies, insofar as they both are “human expressions”  [Busa 1999, 6]. Nevertheless, what do we really know about the language (our language) on which human communication is based and by which it is conveyed? Are we really aware of the intrinsic logic, the inner dynamism, and the psychological implications which are put in motion every time we communicate? Roberto Busa defined the “language that [is] unknown”  [Busa 1999, 6][3], as signifying that establishing an interaction with the computer/machine implies a more profound level of language awareness than that to which we are accustomed. This raising of consciousness is necessary to introduce consolidated philological and linguistic methods to the “new qualitative dimensions”  [Busa 1980] made available by informatics. In the light of this, applying computer methods to the humanities “can help us to be more humanistic than before”  [Busa 1980, 89] because it leads us, first, to an inner journey through rational paths triggered by language expression. This conviction mirrors the current debate regarding the necessity of an epistemological reflection on the making of the digital humanities: for example, Stephen Ramsay and Geoffrey Rockwell have recently stated that “the understanding of underlying theoretical claims is the sine qua non of humanistic enquiry”  [Ramsay and Rockwell 2012]. Roberto Busa would have certainly confirmed this vision and reasserted that these “underlying theoretical claims” primarily involve language dynamisms and linguistic issues. More specifically, these underlying implications concern the meaning, the “semantics” of words and sentences, which is not attainable by a mere quantitative production of a certain amount of data. Busa claimed that “we do not speak in words but in sentences. A sentence has a global meaning which is not the pure sum of the values of its single components. The heart of this problem is whether we are able to formalise the global meaning of sentences with something less than the whole sentence itself; in other words, whether we can succeed in identifying in each sentence something which can be taken as characteristic of its global meaning”  [Busa 1980, 88].

According to Busa’s experience, a high-quality methodology in computing for the humanities should be focused on an accurate reflection on communication, because man and computer interact with the help of sophisticated languages which differ from common grammar and syntax and require constant development. Charles L. Isbell et al. have confirmed the language-oriented nature of the interaction established by human beings with the machine and suggested a reciprocal shaping-power of the language on the outside reality (the computer) and, vice versa, of the external world (the computer) to the language itself [Isbell et al. 2009].[4]. A more conscious remark of this communicative essence may contribute to illuminating a controversial point in computing for the humanities: as Johanna Drucker has pointed out, “the challenge is to shift humanistic study from attention to the effects of technology (from readings of social media, games, narrative, personae, digital texts, images, environments), to a humanistic-informed theory of the making of technology”  [Drucker 2012]. With respect to this question, Busa's message is that the development of an “informed theory of the making of technology” cannot elude a renewed, unremitting consideration of language features and issues, which, in a kind of virtuous circle, are positively related to the effects of technology on our life. In this sense, the fruit sprouting from the “hermeneutic informatics” to which Busa referred in his last keynote speech From Punched Cards to Treebanks mentioned previously is what I propose to indicate as a language-based hermeneutics of informatics.

This type of theoretical approach, along with the scepticism Busa expressed regarding possibilities for a machine to fully develop it [Busa 1990][5], clears an unexpected, irreplaceable space for the human role: in fact, although “the computer has even improved the quality of methods in philological analysis, because its brute physical rigidity demands full accuracy, full completeness, full systematicity”  [Busa 1980, 88], investigating the meaning of a discourse, as well as the mind processes involved, requires that man instructs the machine to work at this level.

Busa found in the so-called arbor Porphyrii (the “Porphyrian tree”, suggested by the ancient philosopher Porphyry to explain the Aristotelian Categories[6]), an effective way of representing how the human way of reasoning works. This is perfectly interpreted by the “trees” which constitute a treebank, a linguistic corpus characterised by a specific focus on the relationships between words and sentences. Busa was aware of the difference among the “Porphyrian trees”, expressing “a graduated scalarity of similarities and differences”  [Busa 2009][7] among genres and species, and the “treebanks”, illustrating “relations of real and true dependency”  [Busa 2009] from a linguistic point of view; nevertheless, he asserts that dependency-trees, provide “a syntax extracted inductively from computerised texts and workable by the computer according to its boundless capacities”  [Busa 2009] that is a pedagogically powerful means for clarifying our interior logic and way of thinking and, consequently, of communicating[8].

Building a treebank for the syntactic and semantic annotation of the 11 million words constituting the Index Thomisticus, as it is currently being performed by the Research Group committed to Busa’s legacy with the Index Thomisticus Treebank Project[9], is thus a method of quality par excellence, not only for the enhancement of linguistic properties in a text but also for casting light on some mental paths connected with language. My current experience as annotator of the IT-TB, in fact, has been increasingly persuading me about what I propose to define as the performative nature of linguistic annotation. While working in Busa’s footsteps, I have been practically experiencing first-hand the insights he advanced in From Punched Cards to Treebanks: 60 Years of Computational Linguistics and, at the same time, I have been realizing how these insights can lead us ahead of what Busa himself could not have completely foreseen. The most intriguing fact of annotation of a treebank is that it enables the human annotator to deal with a living, branching and constantly evolving tree. Making a tree live, by establishing relations between words until it has reached its accomplished form, is the amazing vocation I have discovered, since, in the very act of annotating, I can make the text live. It really represents a kind of “making of language,” to recall the title of the book by Mike Beaken, The Making of Language [Beaken 2011], in which is proposed an alternative view of looking at the origins of language, rooted not so much in biological features, rather than in cultural and technological developments. Treebanks, in my vision, represent one of these contemporary, worth exploring and pursuing developments, made available by the advancements of linguistics not only for modern, but also for no-longer spoken languages.

Language-based models of interpreting and building a language, along with the development of new technologies in the field of computing applied to the humanities, may represent theoretically and practically valid methods of contributing to the above-quoted challenges explained by Unsworth and Davis. Despite being software-embedded, in fact, these models and related technologies may be able to overcome the inherent limitations (e.g., strong object-orientation) of mere computational approaches not capturing the performative value of human acts of linguistic expression. In addition, owing to our language-shaped minds, they may offer a remedy to that “universal lament about the fragmentation of knowledge”  [Busa 2009] that Busa ascertained, because “the human hunger for syntheses derived from microanalysis, continues to surge up, and not only in technologies, but also in linguistics, philosophy, psychology and theology”  [Busa 2009] — in other words, in the context of general human knowledge.