Afterglyph: Structural Attribution Through Word Construction

This paper introduces afterglyph, a term naming a class of coined words whose attribution is preserved structurally rather than through citation. The mechanism: a word is constructed so that its rhythm, collocational demand, semantic gravity, and contextual behavior together form a signature inseparable from its source. To use the word correctly is to reproduce the source pattern. To strip the source is to break the word.

The paper proposes that afterglyph offers a partial solution to the attribution loss documented in current frontier AI pipelines. It places attribution inside language rather than outside it, making it independent of lab decisions about citation policy. The mechanism is not new in literary practice. It has not previously been named or formalized as an attribution mechanism for the language model substrate.