Two lexical decision experiments compared semantic and repetition priming by masked words. Experiment 1 established prime-mask stimulus onset asynchronies (SOAs) with presence-absence detection judgments. Primes presented at detection-threshold SOAs produced equal facilitation for repeated and semantically related targets: 26 ms and 24 ms. Experiment 2 established SOAs with semantic judgments. Primes presented at 70% of the semantic-threshold SOA to mimic the exposure conditions of Experiment 1 produced slightly greater facilitation for repeated targets but a tendency toward inhibition for semantically related targets: 38 ms and -6 ms. These results confirm Dagenbach, Carr, and Wilhelmsen's (1989) report that strategies induced by threshold-setting tasks can influence masked priming. In addition, Experiment 2 suggests a mechanism for retrieving weakly activated semantic codes into consciousness that relies on the center-surround principle to enhance activation of sought-for codes and to inhibit related codes stored nearby in the semantic network.

In response to Greenwald's article on contemporary research on unconscious mental processes, the authors address three issues: (a) the independence of much recent research and theory from psychodynamic formulations; (b) the broad sweep of the psychological unconscious, including implicit perception, memory, thought, learning, and emotion; and (c) the possibility that the analytic power of unconscious processing may depend both on the manner in which mental contents are rendered unconscious and the manner in which they are to be processed.

No abstract

No abstract

Two experiments investigated semantic priming effects in a modified version of the Dagenbach, Carr, and Barnhardt (1990) rare-word paradigm. After learning a list of rare words to a criterion of 50%recall, subjects participated in a lexical decision task in which the rare words served as primes. Whenthe targets were associatively related to the primes, lexical decision responses were facilitated following recalled definitions and inhibited following unrecalled definitions. When the targets were synonyms of the rare words, facilitation occurred following both recalled and unrecalled definitions. The results were interpreted as supporting a center-surround model of attentional retrieval that may serve an adaptive role in new learning.Knowledge concerning the meanings of words has usually been attributed to a functionally distinct system known variously as lexical memory (Oldfield, 1966), generic memory (Hintzman, 1978), or semantic memory (Tulving, 1972). Access to a word's meaning on the basis of its orthographic or phonological code appears to occur rapidly and automatically, and appears also to activate the representations of semantically related words. This latter phenomenon, known as semantic priming (see, e.g., Meyer & Schvaneveldt, 1971), has been studied extensively and has prompted a number of theoretical formulations concerning the organizational structure ofand processes associated with the semantic memory system (see Neely, 1991, for review).In the standard semantic priming paradigm (Neely, 1977), subjects are required to make a lexical decision to the second of two consecutive letter strings. When both strings are words, subjects are faster to respond when the words are semantically related than when they are not. Such semantic priming effects have usually been interpreted within a spreading activation model of semantic memory (Collins & Loftus, 1975). According to this model, concepts are connected by pathways that vary in length according to their degree of semantic relatedness: Highly related concepts are close in semantic space and unrelated concepts are far apart. When any particular concept is activated, activation spreads along the pathways to related nodes such that concepts close in space benefit more than distant concepts. In the semantic priming paradigm, activation spreads from the prime to the related target node, which can then be accessed more quickly so that lexical decisions are speeded. This spread of activation appears to occur automatically. Snyder (1975a, 1975b) proposed, however, that in addition to an automatic spreading activation mechanism, strategic or attentional processes that reflect the expectations or intentions of the subject may also be involved in semantic priming. These theoretical ideas received empirical support in a series of studies by Neely (1977) in which he manipulated the expectations of the subjects and the prime-target stimulus onset asynchronies (SOAs). At short SOAs (i.e., 250 msec), expectations concerning the type of targets had no effect on the stand...