Jean Piaget's ideas regarding symbolic function are expanded in this paper to provide a model to use in distinguishing between general symbolic versus specific linguistic deficits in language disordered children (whose disorders are not due primarily to intellectual, sensory, motor, or social-emotional deficits). In applying this model to the assessment of language disordered children, the paper suggests that nonverbal cognitive assessment needs to distinguish between symbolic imagery and nonrepresentational aspects of nonverbal cognition. The paper also reviews the literature relevant to the issue and notes that, in light of various methodological problems encountered in these studies, the evidence to substantiate the existence of broad representational deficits and specific linguistic deficits in this language disordered population is far from conclusive. The paper concludes that a knowledge of the problems encountered in this previous research should serve to guide future research efforts. (Author/FL)