| Recent investigations in our lab have accumulated data that challenged symbolic theories (for example, rule learning explanations of behavior). Both computational and behavioral work have suggested that subsymbolic theories (for example, associative explanations) may be sufficient to account for serial behavior (Wallace and Fountain, 2000a,b; Stempowski, Carman, and Fountain, 1999). The current set of experiments assessed the utility of subsymbolic explanations of serial behavior in an operant procedure in which rats were required to respond to a series of levers arranged on the walls of an octagonal chamber (Fountain and Rowan, 1995a,b). The sequence of responses may be varied to create sequences that have specific properties. Experiments 1 and 2 examined the extent that a variety of stimuli come to control responding to sequences that can be described by three nested rules, that is, to three-level hierarchical sequences. In Experiment 1, rats first learned the sequence, 123 234 345 456 567 876 765 654 543 432, for 28 days then spatial cues and temporal phrasing cues were manipulated alone and in combination. In Experiment 2, rats first learned the sequence, 12 123 2345 45678 87654 5432 321 21, for 28 days then chunk length and spatial cues were manipulated alone and in combination. After retraining on the original sequence for two days, chunk length plus phrasing cues, spatial cues plus phrasing cues, or all in combination were manipulated.; Cue modification transfers disrupted specific aspects of performance in both experiments. These disruptions indicated that spatial, phrasing, and proprioceptive stimuli controlled some aspects of serial responding in three-level hierarchical sequences. Spatial transfers disrupted responding at the third-level rule transition, indicating that spatial cues controlled responding at this transition. Phrasing cue removal disrupted responding at chunk boundaries, and other evidence indicated that proprioceptive cues guided performance in the absence of phrasing cues.; Experiments 1 and 2 provide evidence that rats use a variety of stimuli to guide performance on three-level hierarchical patterns. Stimulus control of responding conflicts with symbolic explanations of serial behavior and supports subsymbolic explanations. |
