Solid State Poetry is an enquiry into agency, perception, and meaning in an age of machine communication. This series, which originated in 2014, is premised on the algorithmic generation and encoding of poetry into kaleidoscopic patterns. These patterns articulate the verbal structures of each poem, and represent a speculative reworking of the means through which information is captured and conveyed within digital systems.

Poetry, as one of the most sophisticated modes of writing available to human actors, radiant with potential readings, appears incommensurable with schemas designed to compress and encode electrical signals for transmission across noisy channels. Both articulate very different imperatives, with one seeking to reduce the potential for novelty, for error, whereas the other is seeking to generate new vectors of thought and expression. Nevertheless, both also represent specific modes of reworking the normative vectors of written language, albeit for very different audiences, human and machinic. Therefore, while the operations of electronic signal processing manifest far below the thresholds of the human sensorium (hinted only transiently in the form of glitches and errors), it might be speculated how they could be recast as a specifically nonhuman mode of poetic performance—one that can inspire a rethinking about what digital systems represent as expressive agencies.

Although the production techniques behind Solid State Poetry have evolved considerably since its inception, the basic encoding processes have remained similar. Initially, a new poem is generated algorithmically using simple cut-up techniques on a source vocabulary, before analysing the results using another algorithm that assesses their constituent verbal patterns, mapping these visually onto a matrix of Truchet tiles.

In this latest iteration of the series, the generated poems are analysed in terms of their lexical properties and overall sentiment. A series of four keywords are selected from those identified by the system as being semantically important for a given poem, based on an assessed lexical ‘score’. The keyword score values are used to generate a basic ‘motif’ pattern, which is tiled repeatedly across the display grid. This sequence is transformed subsequently by a set of specific rules that are assigned to each motif. The final output is then colourised using a set of fixed palettes, which are selected based on the detected sentiments of the generated poem—as evoking mainly either positive or negative imagery, depending on the vocabulary present.

Although this process has the effect of rendering the poetic source entirely inaccessible to a human observer, the patterns yielded evoke the machinic umwelt that is integral to the operations of computer mediated communication, in which messages are prehended in terms of standard encoding schemas, error-correcting checksums, and varied data analysis metrics. More fundamentally, it signifies how, across the communicative spaces defined by digital infrastructure, the vast majority of signals are designed to be received and interpreted by machines only.

It is with reference to both these points that the asemic ‘silence’ of Solid State Poetry is working to denature the relations between message and receiver, as defined by traditional, humanistic conceptions of aesthetic engagement, and hinting instead at a distinctly machinic mode of reading the spaces and durations of the contemporary electronic environment.

The poems themselves are never revealed to an onlooking audience, who are left free to consider the relations between the visual rhythms they discern with the verbal forms they echo.

Example plates from Solid State Poetry