4.0 The lexicon – as conceived of by linguists – is a central component of natural language. It structures in some way the raw materials that may serve as inputs to various systems of semantics, morphology, syntax, and phonology. This much is generally agreed upon, if only because it skates over much we don’t really understand.
4.1 Much of the power of language lives in its combinatorial capabilities with the materials of the lexicon. In Aspects of the Theory of Syntax (MIT, 1965) Chomsky famously quotes Wilhelm von Humboldt’s assertion that language ‘makes infinite use of finite means,’ and indeed much theoretical linguistics, particularly in the field of syntax, involves positing and refining mechanisms that would account for this apparent state of affairs. Without getting too bogged down in the particulars of controversial territory, we can imagine the skeleton of such a mechanism – one that is able to traverse in some manager the organisation of the lexicon, referencing and selecting elements for composition and utterance. My simple mind reaches for corollaries such as typesetters pulling from cases of metal type for printing, or composite video signals formed of the complex amalgam of many sine waves representing different colour levels. Better still; a Turing machine that reads in its act of selection, but also writes information concerning this latest instance of use, so as to affect future occasions. We must surely imagine that the contents of the lexicon are not islands unto themselves but possess some means of hinting at the ways in which they may readily combine with other elements, and that this latent relationality is built into the structure of the lexicon itself.
4.2 For all poetry’s special attributes, and my own interests in its most extreme forms of artifice, I remain invested in the notion that composing a line of poetry makes use of much the same natural language faculties as does a conventional sentence or utterance. For even the most intense poetic experimentation takes place over the primal ground of language, no matter the degree to which it dislocates us by ostentatiously flouting conventions. If we are to take seriously Reddy’s claim of erasure as a ‘technology of writing’ – as I’m suggesting we should – then we must imagine how it fits into our picture of manufacturing any other kind of language. Instead of a mechanism traversing the lexicon to make selections for combination, we engage in a bit of outsourcing and move through a text making selections for composition. Instead of the relational structure of the lexicon defining potential connections between things, the discursive structure of the text influences the landscape through which our roving mechanism can move, as well as the concepts and textures that may manifest.
4.3 Erasure is a statement about the nature of the lexicon, and how we can experience it to some degree in a phenomenological sense as an at least partly self-contained module that can be changed out for something else – like pointing to a new drive storing different data. This experience touches what is normally a deep and inscrutable part of our own interiority, with mysterious primordial origins in our cultural and physiological evolution. It allows us to play in this depth by bringing something from the outside world into that interior space of our being, to see what will happen and how our faculties may adapt. Perhaps more than any other kind of formal poetic R&D, erasure permits us to interrogate the algorithms that constitute our artistry through controlled experiments with diverse materials; to assess our sampling biases and propensities that are anything but inevitable.