literalto

Literalto is a theoretical framework and encoding standard used in linguistics and artificial intelligence to represent language in its literal, non-figurative form. It aims to convert utterances into a canonical semantic representation that preserves propositional content while minimizing metaphor, idiom, and rhetorical devices.

Origin and scope: The term was introduced by researchers in the LINGO-LITE collaboration during the early 2020s. Literalto defines a three-layer model: surface normalization to a literal base form, a semantic layer built from predicate-argument structures, and an interpretation layer that adjusts readings based on context and domain.

Mechanics: The system relies on a constrained lexicon of core senses and uses frame-based representations with semantic roles such as Agent, Patient, and Instrument. Sentences are first mapped to a canonical proposition, then augmented with modifiers that reflect necessary temporal, aspectual, or evidential information.

Examples: The sentence It's raining cats and dogs is paraphrased as Rain falls from the sky with high intensity. The idiom spill the beans is paraphrased as Beans are spilled by someone, illustrating the literal-to-literal rendering. In practice, idioms may be tagged as non-literal when a literal reading would distort meaning.

Reception and limitations: Critics note that literalto can strip expressive nuance and irony, and may struggle with highly contextual or metaphorical discourse. Proponents argue it supports deterministic parsing, improved translation quality for technical domains, and clearer data for downstream reasoning systems.