By Fiona Baillie and Florencia Pistritto
Andrew Ash: Translating the Present into the Future, Saturday 26 February 16:00
Does a Saturday afternoon spent pondering over word choice and imagery used by T. S. Eliot, Henry James, Keats and Shelley, or the abstract nature of Kandinsky’s art appeal to you? At Scottish CIOL’s February online event, Andrew Ash, doctoral researcher from Alabama University, transported attendees into the world of culture and art. His presentation provided real insight into what makes the work of some artists so unique, and why literary masterpieces, particularly poems, are so difficult to translate. His starting point, bizarrely, was A.I. and word prediction algorithms.
Ash had an ‘aha’ moment when he noticed that the algorithms in popular applications (the ones that try to predict the next few words of the email you are typing) consistently failed to predict the next words of poems, regardless of how many preceding words of the verses he gave them. The poems he studied were written in English, but not the standard English we might use every day. For example, in Keats’ Ode to a Grecian Urn, we see unusual word combinations such as “bride of quietness” and “foster-child of silence”. Algorithms would never connect ‘bride of’ and ‘quietness’ because these words don’t normally go together. Language and predictive algorithms work on the principle that context matters, that what is likely to come next will be a word that is in a group of ‘close friends’ of previous words. This tends to work when we use ordinary language. However, literary language is different: dramatic, abstract, the product of the author’s imagination.
This shows us how poetry creates new concepts and meanings by effectively translating familiar language into new language. This new language might include unusual structures, word combinations, concepts and original language patterns. This produces a distortion of language, a sense of strangeness, which in turn challenges the reader. If a poem requires effort to understand and decode, the reader becomes absorbed in that task. The process is also likely to bring about an emotional response in the reader. After all, art is about communicating feeling, often in a sophisticated and multifaceted manner.
This theory helps explain something that most linguists probably know whether they work as literary translators or not: that translating literature is extremely difficult. It requires the translator to create new linguistic forms in a second language that parallel those created in the original text. It entails the translator thinking about words and experiences, using intuition and understanding the artistic ambition of the original writer.
Andrew Ash ended his talk by describing how the idea of translating existing linguistic forms into new ones is central in other theories, such as Bruno Latour’s theory of how scientific theorizing works and Homi K. Bhabha’s theory of how new cultures are created at the intersection of existing cultures.
The Scottish CIOL committee would like to thank Andrew Ash for his excellent and thought-provoking talk, for explaining complex academic research in an accessible way, and for converting us all into culture vultures for the afternoon.
Certain cultural formations—poetry, scientific theorizing, and the intersection of disparate cultures—involve the reorganization of existing linguistic structures into new concepts and patterns in a way that can usefully be described as the translation of existing language into new language, of the present into the future.
This talk stems from a little experiment I did in which I found that the word prediction algorithms we are all familiar with in Gmail and Outlook—the ones that (often annoyingly) try to predict the next few words of the email you are typing—consistently failed to predict the next words of poems I typed into them, regardless of how many preceding words of the poem I gave them. To my mind, this shows that the words in a poem organize language in a very different way than the concepts of ordinary language. Rather than using existing linguistic structures to represent familiar meanings, the goal of poetry is to create new concepts, new meanings, and it does this by effectively translating familiar language into new language.
Among other things, this theory helps explain why translating literature is so difficult: it requires the translator to create new linguistic forms in a second language parallel those created in the original text.
The talk will conclude by showing that this theory of the way language works in poetry is directly analogous to Bruno Latour’s theory of how scientific theorizing works and to Homi K. Bhabha’s theory of how new cultures are created at the intersection of existing cultures. In each case the idea of translating existing linguistic forms into new ones is central.
Andrew Ash is a doctoral researcher in English Literature at the University of Alabama. His research focus is poetry, poetics and the theory of reading. His dissertation project, entitled Metaphor/Metamorphosis: A Network Theory of Literature, adopts an interdisciplinary approach and argues that works of literature, scientific theorizing and the intersection of disparate cultures resist understanding in terms of existing linguistic concepts, and instead create new concepts by reorganizing the network of relationships among existing linguistic structures, with consequences for our understanding of how the mind, world and language relate to one another. He has recently spoken at conferences held at the University of Glasgow, Carnegie Mellon University, Indiana University, the University of Virginia, and Johns Hopkins University, and on panels organized by the International Congress of Medieval Studies, the German Association for Postcolonial Studies, the Association of Literary Scholars, Critics and Writers, and the Northeast Modern Language Association.
This event will take place via zoom, please book your place here