Object knowledge comprises a virtually limitless feature space. When cued to generate attributes for an exemplar (dog), any feature is possible (has molecules). Yet, some features are definitively favored (tail). We hypothesize a sensorimotor bias in feature generation wherein perceptually salient features upon first pass are more cognitively accessible than abstract, verbally mediated knowledge. We examined the role of sensorimotor dominance in semantic feature generation by yoking concreteness values to cues (N=4436) and features (N=69,284) within the Buchanan et al. (2019) norms. We predict that cues regardless of their concreteness evoke relatively more concrete features (e.g., dogs evoke tails, justice evokes lawyers). The data moderately supported this hypothesis. Feature concreteness increased linearly with cue concreteness (R=.83) but the y-intercept (2.78) indicates that overall, features were more concrete than their cues. We discuss alternate factors (e.g., frequency, familiarity) that may moderate the likelihood that people retrieve tail when cued with dog.