A set of properties is proposed as characteristic of main verbs, able to identify a particular constituent in a sentence as its main verb. These properties are in turn applied to the framing typology of , in which languages fall into two categories based on whether they characteristically express the Path component of a Motion event in the verb (verb-framed languages) or in the satellite and/or preposition (satellite-framed languages). later held that certain languages do not fall neatly into either category of this typology, either because the lexical category of the Path constituent is unclear, or because the Path and coevent constituents are both verbs or both satellites. He classed such languages together as being “equipollently framed”, exhibiting a third category of the typology. We dispute most of these claims of equipollent framing, demonstrating how these languages are actually either verb-framed or satellite-framed.
For early pre-language hominins, the vocal-auditory channel of communication as then organized may have been unable to accommodate any enhancement in the transmission of conceptual content due to three limitations: comparatively low degrees of parameter diversity, iconicity, and fidelity. We propose that these limitations were overcome by an evolutionary development that enabled an advance from the fixed holophrastic calls of earlier species to the open-ended spoken language of our own species. What developed was a “combinant” form of organization.
Such combinance is a system in which smaller units combine to form larger units. At its smallest scale, this process yields a “clave”. In a clave, generally, units from an inventory at a lower tier combine to form the units of an inventory at the next higher tier in accord with a particular set of constraints. In turn, such claves function as the smaller units that combine to form a larger unit, a concatenation, where the higher tier of one clave serves as the lower tier of the next. The longest such concatenation in language consists of six successive claves. Phonetic features combine to form phonemes under the constraints of feature assembly; phonemes combine to form morphemes under the constraints of phonotactics; morphemes combine to form complex words under the constraints of morphology; morphemes and complex words combine to form expressions under the constraints of syntax; expressions combine to form a single speaker’s “monolog” under the constraints of discourse rules; and such monologs combine to form an exchange between speakers under the constraints of turn-taking.
Our analysis characterizes communication at its most general and contrasts different channels of communication. In particular, the vocal-auditory channel of spoken language is extensively contrasted with the somatic-visual channel of signed language, whose classifier system largely lacks the three limitations of the former. To show this difference, the limitations are analyzed in detail (e.g., iconicity is shown to be based on six properties: prorepresentation, covariation, proportionality, proportional directness, cogranularity, and codomainality). In accord with this difference, the signed classifier system demonstrates the cognitive feasibility of communicating advanced conceptual content with little combinance, but the vocal-auditory channel is seen to have needed the incorporation of combinance for spoken language to evolve.
If a language regularly places a particular closed class in syntactic construction with an open class—for example, nominal affixes with noun roots, or satellites with verbs—any conceptual category expressed by the closed class tends not to be expressed by the open class. This proposed tendency is here called “semantic unilocality”.
Cognitively, semantic unilocality in language may arise from several more general tendencies, including one to avoid redundancy and one to segregate the representation of different conceptual categories. Both of these may in turn arise from a still more general tendency toward communicative efficiency.
Historically, the rise, extended presence, decline, or extended absence of a closed class that expresses a particular conceptual category may foster certain diachronic processes in a syntactically associated open class. These processes include leaching, culling, shift, filtering, and abstention.
This article proposes how language relates structurally and evolutionarily to cognition. It heuristically divides cognition into cognitive systems and the organizing factors that structure them. The general finding is that cognitive systems share these structural properties to different degrees. This is termed the “overlapping systems model of cognitive organization”. The specific finding is that the cognitive system of language shares many structural properties with the cognitive systems of visual perception, of somatosensory perception and motor control, and of understanding, but shares few structural properties with those of affect and of culture.