AbstractNew data on the large number of modality-specific areas in thepost-central cortex of several non-human primates, and recentanatomical and functional studies of the human brain suggestthat very little of the cortex consists of poly-modal 'association'areas. These observations are used to reinterpret psychologicaland neuropsychological data on language comprehension in nor-mal and brain-damaged humans. I argue that languagecomprehension in sighted people might best be thought of as akind of code-directed scene comprehension that draws heavilyupon specifically visual, and probably largely prelinguistic pro-cessing constraints. The key processes of word-recognition andthe assembly of visual word meaning patterns into interactingchains, however, may be mediated in part by species-specific ac-tivity patterns in secondary auditory cortex similar to thosegenerated by uninterpreted speech-sound sequences.One obvious reason to study non-human primate brains is thatthey resemble the human primate brain in many ways. Yet hu-mans exhibit behaviors--especially the comprehension oflinguistic discourse--that are qualitatively very different frombehaviors of primates and other animals. Because of this, somehave concluded that animal brains may be poor models for thehuman brain. There are presently quite substantial rifts be-tween psychological, neuropsychological, and neurobiologicalapproaches to language. Recent developments in studying hu-man and animal brains, however, provide a strong impetus tore-open discourse among these disciplines.The neocortex of all mammals is now known to consist pri-marily of a mosaic of visual, auditory, somatosensory, motor,and limbic areas. Primitive mammals have a small number ofareas in each of these modalities, while carnivores and primateshave many. In monkeys, for example, a mosaic of 25 visual ar-eas occupies more than half of the entire neocortex (Merzenich& Kaas, 1980; Sereno, 1988; Felleman & Van Essen, 1991;Sereno and Allman, 1991). The traditional site for higher-levelfunctions--"polymodal association cortex"--has been reducedto a few diminutive strips in between large expanses of unimo-dal visual, auditory, and somatosensory areas. The potentialsignificance of this reparcellation of cortex for the study of lan-guage and the brain has hardly been explored. The aim of thispaper is to re-introduce a thoroughly comparative perspectiveinto the evolutionary acquisition of the capacity for language,but one that does not back away from the obvious cognitive dif-ferences between humans and other animals. The anatomicaland physiological organization of cortical areas in primates, in-cluding recent work on human cortex, is reviewed first. Theimplications of this work for theories of human language com-prehension are then explored.Cortical Sensory Areas in PrimatesDefinition of a Visual Area. Cortical sensory areas are bestdefined by multiple converging criteria (Van Essen, 1985;Sereno and Allman, 1991). I begin here with visual areas, sincethey constitute the largest of the primary subdivisions of thecortex. Criteria for the definition of a visual area presently in-clude architectonic features (e.g., degree of myelination, cellsize, cell morphology, and cell packing density in cortical lay-ers, histochemical features), connection patterns (e.g., inputand output areas, laminar origins and targets of connections),visuotopic organization (e.g., mirror-image or non-mirror-im-age map of hemifield, bounding areas, pattern of mapdiscontinuities, degree of retinotopy), and physiological prop-erties (e.g., excitatory receptive field size, direction selectivity,attention-related modulation). Areas differ in the degree towhich these criteria have been explored. V1 (primary visualcortex) and MT (middle temporal area) are distinct, well-stud-ied areas in primates that are convergently identified by manyof these criteria. Other areas--e.g., in inferotemporal cortex--are less well studied. There is no evidence to suggest that theyare any less distinct.Visual Areas in Prosimians and Monkeys. The first primateswere probably nocturnal, judging from the large size of their or-bits. The primates living today most closely related to theseearly primates are also nocturnal or crepuscular. The bush babyor galago, the only prosimian primate studied in detail (Allmanand McGuinness, 1983), has on the order of 16 visual areas.Almost all visual areas in galagos exhibit a substantial degreeof retinotopic organization, including areas in the inferotempo-ral cortex. In these studies, the entire extent of visual cortexwas physiologically mapped in detail for the first time. In apassive animal, visual areas only respond to visual stimuli, au-ditory areas only to auditory stimuli, and somatosensory areasonly to somatosensory stimuli. Visual cortical areas border al-most directly upon somatosensory areas (dorsally) and auditoryareas (ventrally). The transitional strip between, for example,auditory and visual areas (in which neurons have both a visualand an auditory receptive field) is less than one mm wide.Monkeys (anthropoids) are thought to have diverged fromthe ancestors of galagos at least 40 million years ago. All butone of the anthropoids are diurnal (day-living), suggestingstrongly that day-living habits evolved early in the monkey lin-eage. The one nocturnal monkey, the New World owl monkey,lacks a tapetum, suggesting that its ancestors had diurnal habits.The organization of visual cortex has been studied in detail intwo different monkeys--the owl monkey and the macaque mon-key. Figure 1B shows a flattened summary map of visual areasin the owl monkey (Weller and Kaas, 1987; Sereno andAllman, 1991). As in galagos, V1 is the largest area, followedby V2. There appear to be at least three somewhat separate’streams’ of information passing through V1 and V2--the mag-nocellular, parvocellular interblob, and parvocellular blobstreams (named after their relay structures in the dorsal lateralgeniculate nucleus and area V1)--that remain somewhat sepa-rated as one moves on to higher areas (Livingstone and Hubel,1984; DeYoe and Van Essen, 1988). These pathways processdifferent aspects of the visual signal in parallel--roughly, mo-tion, location, and depth in the magnocellular pathway, andcolor, shape, and shading in the parvocellular pathways. Thepathways pass through layer 4B, layer 2-3 interblobs, and layer2-3 blobs in V1, and the thick stripes,