The automatic processing hypothesis holds that very familiar items are processed involuntarily, and two color-naming experiments were designed to test this further. Experiment 1 employed words as written stimuli and focal and nonfocal colors as inks. Supporting the automatic processing hypothesis, neutral words delayed color naming, indicating that their representations were activated involuntarily. There was no interaction of ink type and written item conditions. Experiment 2 employed single letters as written stimuli. Color-name initials delayed color naming when incompatible with the color and facilitated it when compatible, suggesting that, although letters may be processed involuntarily, the way in which they are encoded is influenced by the cognitive context. The data from both experiments were discussed in relation to current interpretations of Stroop effects.Several theorists have suggested that there are two fundamentally different modes of information processing, automatic and nonautomatic (LaBerge & Samuels, 1974;Posner & Snyder, 1975; Schneider & Shiffrin, 1977). They describe automatic processing as developing with practice, occurring involuntarily, and being independent of any limited capacity central mechanism. An example of such automatic processing is reading a word. For the child, reading is initially laborious and slow, and may proceed on a letter-byletter basis. For the adult skilled reader, reading requires little effort, is fast, and can be performed concurrently with other activities, as when one reads a billboard while driving. In reading a word, the automatic processing theorists suggest that an internal memorial representation of the word is automatically activated when the word is presented, and that such a representation contains information about the word's pronunciation, meaning, and association to other concepts.The focus of the present experiments is on the involuntary nature of automatic processing. Experiment 1 examined the involuntary processing of words. Experiment 2 employed single letters. Theories of automatic processing suggest that people have no direct control over automatic processes, that no decision making is involved, and that the processing sequence will run to completion whether or not it This paper was written while the author was at NASA, Ames Research Center, Moffett Field, California, under a National Research Council Research Associateship. I would like to express thanks to Eleanor Rosch, who advised on the experiments and commented on the article, and to Stephen Palmer and Loy Braley, who commented on earlier versions. Requests for reprints should be sent to Joan Regan, Department of Psychology, University of North Carolina, Greensboro, North Carolina 27412.is compatible with the person's intention. Posner and Snyder (1975) have indicated that the Stroop phenomenon is a prime example of automatic processing without intention.The Stroop task requires people to name the color ink in which a word is printed as quickly as they can. This task is difficult to do w...

These experiments attempted to confirm the description of automatic processing using a naming task. Automatic processing has been described as developing with learning, being attention free, and being involuntary. Experiment 1 employed a naming task and addressed the attentional capacity issue by testing performance in pure and mixed-list conditions while varying the number of items in the lists. Armenian letters showed strong effects of processing load, and this decreased with increasing familiarity. Performance with English letters in the pure lists showed a considerably lesser effect. Experiment 2 investigated the relationship between familiarity and involuntariness in processing. Subjects named small letters, several of which combined to form a large letter. Both large English and large Armenian letters delayed naming small letters, although when the large letter was the same as the one to be named there was no interference. Experiment 3 replicated this finding for naming small English letters, using subjects totally unfamiliar with Armenian letters. Again, both sets of letters delayed naming. The data are in agreement with Navon's global-processing hypothesis but are contrary to standard response-conflict interpretations of the Stroop effect. Experiment 4 compared performance with the two sets of letters in a physical matching task to rule out perceptual complexity or interitem confusability as confounding factors in the previous experiments. There were no differences in matching English and Armenian letters. Overall, the data indicate that extensive practice may be a necessary condition for capacity-free processing but may not be a necessary condition for involuntary processing.The adjective automatic has been used to 1967), activation of memory structures (Posdescribe several different sorts of cognitive ner, 1973; Posner & Snyder, 1975a), cornoperations. The description may refer to an ponent stages of a particular task (Logan, entire class of operations, for example, preat-1978), or some task taken as a whole tentive processes (Keren, 1976; Neisser, (Schneider & Shiffrin, 1977;Shiffrin & Schneider, 1977). The exact sense in which Experiments 1 and 2 formed part of a dissertation thg word " automat i c » j s use d also varies.

No abstract

The effect of stimulus familiarity on physical matching was examined in three experiments. In Experiment 1, subjects matched English and Armenian letters. String length varied from one to three items. Familiarity had no effect when two single items were compared, but it had increasingly marked effects with two- and three-item strings. The pattern of same/different responses and serial position effects implicated the comparison process for three-item matches. Experiment 2 varied stimulus onset asynchrony to assess the contribution of familiarity to speeding up encoding and/or to maintaining a representation in memory; only two-item strings were used. Encoding Armenian letters required more time than encoding English, but there was no additional decrement due to holding the representation in memory. In Experiment 3, one group of subject learned names for the Armenian letters and another group practiced drawing the letters. The availability of names did not decrease the familiarity effect. These data suggest that familiarity invariably has a strong effect on the encoding process and influences comparison processes when task requirements are sufficiently demanding.

Acquisition and retention of a list of paired associates were measured either by cued recall or by multiple-choice recognition. The method of testing used during original learning was combined factorially with the type of test administered after a 1-week interval. Speed of learning to criterion under the two procedures was comparable. Long-term recognition was substantially higher than recall. This difference did not interact significantly with the method of learning. Recall benefited substantially from a prior test of recognition, but there was little effect when the order of tests was reversed. It is concluded that (a) both occurrence and retrieval information were stored under each condition of learning, and (b) occurrence information is less subject to forgetting than retrieval information.