|UFDC Home||myUFDC Home | Help|
This item has the following downloads:
1 THEMATIC ROLE PRIMING OF RELATED VERBS: EFFECTS OF MULTIPLE PRIMES By STACEY HERLOFSKY A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MA STER OF ARTS UNIVERSITY OF FLORIDA 2011
2 2011 Stacey Herlofsky
3 ACKNOWLEDGMENTS I am grateful for all the support I ha ve received over the course of completing this project. I appreciate all the advice and guidance that my committee has prov ided throughout this process, as well as the invaluable criticism that has helped me to develop my ideas. I want to thank the members of the University of Florida Aphasia and Bilingualism Lab, where I was fortunate to be offered facilities and assi stance in conducting this study. I am especially grateful to Dr. Lisa A. Edmonds who has been a mentor to me in addition to serving as chair of this committee, Vanessa Frieler, who conducted the stimuli development questionnaire that provided me with a pool of potential stimuli, Katie Austin and Tabitha Maldonado, who acted as experimenters in the experimental portion of this study and the participants who made it possible. I can only hope that this experience has been as valuable to the many people who contri buted as it has been to me.
4 TABLE OF CONTENTS page ACKNOWLEDGMENTS ...............................................................................................................3 LIST OF TABLES ...........................................................................................................................6 LIST OF FIGURES .........................................................................................................................7 LIST OF ABBREVIATIONS ..........................................................................................................8 ABSTRACT ...................................................................................................................................10 CHAPTER 1 BACKGROUND AND INTRODUCTION ...........................................................................12 Semantic Priming Effects .......................................................................................................12 Thematic Roles ................................................................................................................16 Multiple Primes ...............................................................................................................18 Continuous List Priming ..................................................................................................21 The Present Study ...................................................................................................................23 Research Questions .........................................................................................................24 Predictions .......................................................................................................................24 2 METHOD ...............................................................................................................................28 Participants .............................................................................................................................28 Population and Criteria for Inclusion ..............................................................................28 Recruitment of Participants .............................................................................................28 Stimuli .....................................................................................................................................28 Stimuli Development .......................................................................................................28 Stimuli Questionnaire ......................................................................................................30 Relatedness Survey ..........................................................................................................33 Stimuli List ......................................................................................................................35 Experimental s timuli ................................................................................................36 Filler w ords ..............................................................................................................36 Nonwords .................................................................................................................37 Apparatus ................................................................................................................................37 Procedures ...............................................................................................................................38 Design and Analyses ...............................................................................................................40 Participant Analysis .........................................................................................................40 Item Analysis. ..................................................................................................................41
5 3 RESULTS ...............................................................................................................................45 Participant Analysis ................................................................................................................45 Item Ana lysis ..........................................................................................................................47 4 DISCUSSION .........................................................................................................................51 Research Question 1 ...............................................................................................................51 Research Questio n 2 ...............................................................................................................52 Research Question 3 ...............................................................................................................53 General Discussion .................................................................................................................54 Future Directions and Clinical Implications ...........................................................................56 APPENDIX A LIST OF EXPERIMENTAL STIMULI .................................................................................59 B RELATEDNESS RATINGS FOR EXPERIMENTAL TRIADS ..........................................60 C PSYCHOLINGUISTIC DATA ON STIMULI ITEMS .........................................................63 D COMPLETE TEXT OF EXPERIMENT INSTRUCTIONS ..................................................69 LIST OF REFERENCES ...............................................................................................................71 BIOGRAPHICAL SKETCH .........................................................................................................75
6 LIST OF TABLES Table page 11 Sample stimuli and experimental conditions .....................................................................26 12 Planned mean comparisons ................................................................................................26 21 Mean ratings, standard deviati ons, and ranges on a relatedness judgment survey by condition ............................................................................................................................42 22 Mean comparisons of relatedness judgment ratings by condition .....................................42 23 Length of letter strings in priming experiment ..................................................................42 24 Group means for controlled psycholinguistic variables .....................................................43 25 Sta tistical comparisons for controlled psycholinguistic variables. ....................................43 31 Mean reaction times by condition for participant results ...................................................49 32 Condition comparisons for participant results ...................................................................49 33 Average reaction times by condition for item results ........................................................49 34 Condition comparisons for item results .............................................................................49
7 LIST OF FIGURES Figure page 11 Continuous list lexical decision task schematic .................................................................27 21 Stimuli development questionnaire ....................................................................................44 31 Graphs of participant and item means ...............................................................................50 32 Graphs of mean differences for each research question ....................................................50
8 LIST OF ABBREVIATION S ANOVA Analysis of variance is a statistical method used to compare the variance of two or more group means. AOA Age of acquisition, from G ilhooly & Logies (1980) normative data, mathematically transformed to match the 100 700 point range used for the other rated psycholinguistic variables. AP Related agent related patient condition in the relatedness judgment survey. AU Related agent unrelated patient condition in the relatedness judgment survey. CNC Concreteness rating with a range of 100 700, derived by merging normative data from three sources, the Paivio Norms (an unpublished expansion of Paivio Yuille & Madigan, 1968), Toglia and Batti g (1978) and Gilhooly and Logie (1980) by a method described by Coltheart (1981). FAM Printed familiarity rating with a range of 100 700, derived by merging normative data from three sources, the Paivio Norms (an unpublished expansion of Paivio Yuille & Madigan, 1968), Toglia and Battig (1978) and Gilhooly and Logie (1980) by a method described by Coltheart (1981). IMG Imagability rating with a range of 100 700, derived by merging normative data from three sources, the Paivio Norms (an unpublished expansi on of Paivio Yuille & Madigan, 1968), Toglia and Battig (1978) and Gilhooly and Logie (1980) by a method described by Coltheart (1981). IRB Institutional Review Board KFFRQ Computational analy sis of present day American English. This number represents the frequency of occurrence of English words, taken from a text corpus of 1 million words. LDT Lexical decision tasks are behavioral priming tasks in which participants are asked to read letter s trings that are either words or readable nonwords and make a yes or no decision about whether each string is or is not a real word. N Number of participants or items included in a mean or analysis NLET Number of letters in words and nonwords NPHN Number of phonemes in words and nonwords
9 NSYL Number of syllables in words and nonwords ns Nonsignificant difference RR Related related condition in the priming experiment. RU Related unrelated condition in the priming experiment. RSI Response to stimulus interval is the time between a behavioral response and the next stimulus in a priming study design. RT Reaction time is the decision latency of behavioral responses. In the present study, RT refers to lexical decision time, recorded in milliseconds. RQ Research qu estion is abbreviated in methods and results to identify which of the research questions are addressed. SD Standard deviation sig Significant difference UU Unrelated unrelated condition in the priming experiment and the relatedness judgment survey. UP Unre lated agent related patient condition in the relatedness judgment survey. UR Unrelated related condition in the priming experiment.
10 Abstract of Thesis Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Arts THEMATIC ROLE PRIMING OF RELATED VERBS: EFFECTS OF MULTIPLE PRIMES By Stacey Herlofsky May 2011 Chair: Lisa A. Edmonds Major: Communication Sciences and Disorders Background Priming tasks are frequent ly used to investigate factors influencing processing of words. Extensive evidence has shown that presentation of a word (target) following a related word (prime) results in faster reaction times compared to unrelated words. Several studies have observed greater, or additive, priming effects of multiple related primes compared to single related primes. This study uses a continuous list lexical decision task, in which targets of interest are unknown to the participants, to test the hypothesis that task sp ecific attentional processing may be a factor in previously observed additive priming from explicitly presented paired primes and targets. Methods A repeated measures design with four experimental conditions, relatedrelated [RR], unrelated related [UR ], related unrelated [RU], and unrelatedunrelated primes [UU] was used. Experimental stimuli were nounnoun verb triads filling related agentpatient action roles in situation schemas (e.g. thiefmoney steal ) and unrelated role fillers (nouns). Experime ntal stimuli were combined with nonwords and filler words in a continuous list. Participants indicated whether each item was a real word (lexical decision). Measures were accuracy and reaction times (RT). Effects of relatedness were examined using repeated measures ANOVA,
11 and post hoc analysis was used to compare mean differences between RR and all other conditions. Results Repeated measures ANOVA showed a significant effect of relatedness across participants (F(3 ,39) = 4.623, p = .004) and items (F(3,29 ) = 3.667, p = .015). Post hoc analyses showed significant differences between RR UU for participant ( p = .021) and item ( p = .011) comparisons and between RRRU ( p = .012/ p = .006), with no significant difference between RR UR ( p = .570/ p = .107). Conclusio ns These results replicate the basic finding that multiple nouns prime related verbs (producing faster reactions) using a different behavioral task (Khader, Scherag, Streb, & Rsler 2003). However, different from previous work, our findings do not sugges t an additive effect for two related nouns versus one. The absence of additive facilitation suggests that attentional processes may have been a component of more robust effects in previous research, as compared to the continuous priming paradigm used in the current study. Our understanding of the semantic relationships evaluated in this study would benefit from further research with more complex verb representations, different thematic roles, the addition of a neutral condition, and in other populations (e .g., older adults).
12 CHAPTER 1 BACKGROUND AND INTRODUCTION Semantic Priming Effects Historically, neural processing of words and language has been investigated in psychological and linguistic studies using a wide variety of behavioral tasks in which part icipants are asked to respond in some way to stimuli that are experimentally varied Stimulus onset asynchrony [SOA] priming experiments have been used to investigate semantic priming effects by using two basic steps; first, using a stimulus known as a pr ime to establish a semantic context, then presenting a target and measuring the accuracy and time of the response, known as response latency or reaction time [RT] (Neely, 1991). Designs of semantic priming tasks are manipulated to investigate factors that influence processing of target words or language structures by contributing to either faster or slower reactions. SOA is a variable used to manipulate the likelihood of conscious processing of primes, referring to the time between the initial presentation of a stimulus (prime) and the onset of the target, and can range from very brief (60ms or less) to a second or more (Jones, 2010; Perea & Rosa, 2002). Although natural language expresses complex interactions of concepts and ideas, most semantic priming paradigms use simplified language tasks in an attempt to isolate certain aspects of language processing. Semantic priming or facilitation occurs when related primes contribute to faster reactions to targets (as compared to neutral or unrelated primes ), w hile slowe d reactions are attributed to interference effect s In an extensive body of research using a single word semantic priming paradigm, presentation of a word following a semantically related word results in faster reaction times compared to unrelat ed or less related primes (Ferretti, McRae, & Hatherell, 2001; Hare, Jones, Thomson, Kelly, & McRae, 2009; McRae, Hare, Elman & Ferretti, 2005; Neely, 1991; Perea & Rosa, 2002).
13 Semantic priming studies have investigated various hypotheses about the neur al organization and processing of language, including types of semantic relatedness, factors in language processing, and the time course of activation and decay of various types of information that may be made available when lexical entries are processed ( Balota, Black, & Cheney, 1992; McRae & Boisvert, 1998; Moss, Ostrin, Tyler, & MarslenWilson, 1995; Neely, 1991; Perea & Rosa, 2002). Semantic priming effects have been used as evidentiary support for hypotheses of a neural basis for many types of semanti c relatedness including synonyms and antonyms; category names and exemplars (i.e. bird, robin); prototypical and atypical category exemplars (i.e. robin, penguin); closely associated words that are either semantically related (i.e. doctor, nurse ) or unrela ted (i.e. bed, pan); verbs sharing qualities related to motor planning and execution (i.e. drop, toss ); as well as mediated associates that are related through a third concept (i.e. lion (tiger) stripes) (Jones, 2010; Moss et al., 1995; Neely, 1991; Perea & Rosa, 2002; Raczaszek Leonardi, Shapiro, Tuller, & Kelso, 2008; Setola & Reilly, 2005). It has also been hypothesized that the degree of relatedness between words has an influence on word processing. Vigliocco, Vinson, Lewis, and Garrett (2004) compare d concrete noun noun related pairs across very close, close, medium and far (e.g. dagger sword razor hammer tongue ) semantic distances (as defined by the number of shared features) using a standard lexical decision task (LDT) to test a hypothesis that conc eptual feature representations are part of the stored meaning of words, and found that RTs increase as semantic distance increases. Task complexity has also been manipulated to investigate the effects of cognitive demands on language processing (Balota & P aul, 1996; Ferretti et al., 2001; Grondin, Lupker, & McRae, 2009). Pronunciation tasks, in which participants are asked to read primes silently, then pronounce targets aloud, require perhaps the least conscious processing of any single word
14 priming paradi gm. In terms of the mean difference in milliseconds between related and unrelated conditions, primed single word pronunciation tasks have produced semantic priming effects that are similar but frequently smaller than more demanding behavioral tasks. Prim ing effects of pronunciation tasks are argued to primarily reflect preactivation due to automatic semantic processing (Neely, 1991; Balota et al., 1992). In lexical decision tasks (LDT), participants are asked to read target letter strings that are either real English words or pronounceable nonwords (i.e. pess ), make yes or no decisions on each target, and respond with key presses. Lexical decisions are more sensitive to semantic relatedness than are pronunciation tasks, but are still argued to be pri marily influenced by automatic processing (Neely, 1991). Semantic decision tasks are similar to lexical decisions in requiring yes or no responses to each target, but semantic decisions require explicit processing of semantic information such as conc reteness ( i.e. Is the target something you can touch? ), or animacy ( i.e. Is the target alive? ), meaning that conscious attentional factors are involved (Hare et al., 2009). Increasing the amount of semantic information that must be accessed by partic ipants to complete a task is a manipulation intended to increase sensitivity to specific aspects of word meaning in semantic priming effects (Grondin et al., 2009). Conscious processing of prime target relationships can also be deliberately maximized, as in relatedness judgment tasks in which participants are explicitly asked to decide whether each target is related to a preceding prime or primes (Kounios, Green, Payne, Fleck, Grondin, & McRae, 2009). In all of the semantic priming paradigms previously de scribed, attempts are made to disguise the prime target relationships of interest by including filler trials that do not have the semantic relationship as the experimental primes and targets, in addition to the nonword or other no response trials. Tasks that require conscious semantic computations such as animacy or relatedness judgment are argued to reflect
15 more attentional processing, compared to relatively automatic semantic processing effects seen in pronunciation tasks or LDT (Ferretti et al., 2001; McRae & Boisvert, 1998). Various forms of distracters have been used to demonstrate an interference effect, in which reaction times are slowed compared to a neutral or unrelated condition. In picture word interference tasks, participants are asked to name pictures as quickly and accurately as possible while ignoring visually presented distractor words, and closer semantic relationships between targets and distracters have been observed to slow reaction times for both nouns and verbs (Sailor, Brooks, Brue ning, Seiger Gardner & Guterman, 2008; Vigliocco et al., 2003). Blocked semantic naming paradigms manipulate the presentation order in rapid naming tasks to establish a semantic context, such that blocks of successive items are either semantically related or mixed, with items cycling through multiple repetitions (Schnur, Schwartz, Brecher & Hodgson, 2006; Vigliocco et al., 2003). Schnur et al. observed an interference effect of increased exposure to nouns within the same category (e.g. animals, appliance s, etc.), with significantly slower reactions across successive cycles for homogonous blocks of category exemplar nouns compared to the same nouns in mixed category lists. Interference effects resulting from activation of multiple, semantically related wo rds or concepts has implications for the understanding of language processing that are not apparent in the single word semantic priming paradigms described above. The present study will employ a modified priming paradigm that attempts to address this issu e. Much of the early research on semantic priming of words focused on the relationships between concrete animate or inanimate nouns, which can be easily and intuitively grouped into superordinate categories (i.e. robin and bluejay share the superordinate category birds ) (Neely, 1991). Expanding from that basis, similar semantic priming tasks have been used to extend our
16 understanding of relatedness. For example, Moss et al. (1995, experiment 1) included prime target pairs with different relation types (category coordinates, i.e. pighorse and functionally related pairs, i.e. restaurant wine ) in a single experimental task and found overall priming effects, with no significant effect of relation type, providing evidence that the contextually related pai rs were related similar to the previously established category relationships. Thematic Roles T hematic roles represents another aspect of the relationship between words that has been tested using semantic priming methods. Thematic roles are components of lexical entries that relax the historical distinction between semantic and syntactic information Both semantic priming and sentence processing studies have provided support for an account of language processing in which the stored meaning of verb s conta in information about the entities that are centr al to the representation of action concept s and that interact with their stored syntactic argument structure s (Feretti et al., 2001; Shapiro, Brookins, Gordon, & Nagel, 1991) Thematic roles include the initiator of action (agent), recipient of action (patient), location, instrument, and event. For example, the meaning of the verb arrest cannot be represented without reference to a person who arrests (agent), and a person or entity that is arrested (patient). The basic syntactic argument for arrest requires subject and object slots that are filled by thematic role nouns to construct a meaningful message. Verbs can be described in terms of their required or permissible thematic structure in addition to information such as syntactic category and argument structure (Shapiro et al. 1991). Extending the theory of semantic relatedness to include thematic roles necessitates the addition of an event memory account, in which real world knowledge and experience about common situations and events forms part of the stored meaning of words (Ferretti et al ., 2001). Thematic role fillers that appear together in a common or prototypical group are referred to as a
17 situation model, or schema (McRae, Hare, Elman, & Ferretti, 2005). Verbs are crucial to the organization of situation schemas, as they carry the representation of the action. Returning to the example of accuse, an event memory account would include experiential knowledge about the types of entities that accuse or are accused in the stored concept of the verb. Research investigating ongoing sentence processing supports the theory that thematic information is part of the lexical entry for verbs (Ahrens & Swinney, 1995; Shapiro et al., 1991). The complexity of sto red representations may be a factor in the activation of verbs. Ahrens (2003) included verbs that demand three thematic roles (e.g. give, which in its most common representation requires a giver, givee, and thing given) and two role verbs (e.g. throw wh ich can include a recipient but only requires a thrower and a thing thrown) in a sentencecompletion lexical decision task in which other potential factors were controlled (a dual task interference method), and observed slower reaction times for three role verbs. Thematic role investigations offer insights into the interactions between the semantic and syntactic information that is made available during activation of verbs. Priming research using both word and sentence contexts suggest that the thematic role information contributing to language processing is not limited to thematic role structure, but includes conceptual information about prototypical or common fillers for those roles. Ferretti et al. (2001, experiments 1 and 4) used a sentence fragment condition, with a cross modal naming task designed to cue specific thematic roles, to compare agents and patients that had previously shown priming effects on related verbs in a single word semantic decision task. Presenting target agents and patients in their congruent and incongruent roles (i.e. He was arrested by the [ cop ], [ crook ]), Ferretti et al. (2001) observed a significant priming effect for both agents and patients in their congruent but not in their incongruent thematic role assignments.
18 Singl e word semantic priming paradigms have been used to explore relationships between nounnoun and verb noun related pairs with specific thematic role relationships, stripped of any syntactic context that could establish expectations for specific thematic rol es. McRae et al. (2005) observed priming effects of agent, patient, instrument, and location noun primes on related verb targets using a pronunciation task. Verbs have been observed to prime noun targets related via situation schemas filling agent, patie nt, and instrument thematic roles using LDT and semantic decision tasks (Edmonds & Mizrahi, submitted; Ferretti et al., 2001). Multiple Primes Effects of multiple primes on targets have also been explored, to evaluate the possibility of additive facilitation for two related noun primes as compared to single related primes (paired with an unrelated prime in the first or second position so that the task is consistent) (Balota & Paul, 1996; Khader, Scherag, Streb, & Rsler 2003). In this paper, the term a dditive facilitation is used to refer to priming effects of multiple primes on targets that are greater than the effects of single related primes combined with unrelated primes in the same task design. Balota & Paul (1996) investigated effects of two noun primes on noun targets with several manipulations of relatedness type and task complexity. Taking advantage of the fact that English allows lexical ambiguities in which entries can have multiple conceptual representations (i.e. organ=
19 LDT and a relatedness judgment task, additive priming effects were seen for targets following two related primes compared to a sin gle related prime in either the first or the second prime position, regardless of the type of relatedness. However, the re latedness manipulation did not a ffect the size of the additive priming effect in the LDT task, while in the relatedness judgment task unambiguous targets received a much larger boost in additive priming from a second related prime. Relatedness judgment tasks demand more explicit semantic processing than LDT, and Balota and Paul (1996) also used different SOAs (299ms in the LDT task and 533ms in the relatedness judgment task), giving participants more time for conscious processing of primes in the relatedness task. In the LDT condition, Balota & Paul (1996) observed similar additive priming effects ranging from 10ms to 33ms compared to single related primes in the first and second positions, across all three relatedness types. In the relatedness judgment task condition, Balota & Paul ( 1996, experiment 6) observed additive facilitation similar to that seen in the earlier experiments for ambiguous targets (16ms faster compared to a single related prime in the first position; 37ms faster compared to a related prime in the second position), while for unambiguous targets additive facilitation was far greater, with a more than 120ms increase in the priming effect compared t o single related primes in either position. Khader et al. (2003) also used a relatedness judgment task to investigate priming effects of multiple primes, using noun and verb stimuli. The overall SOA for each trial in Khader et al. (2003) from presentation of the first prime to onset of the target, was 1300ms. Prime target relationships were not strictly controlled for thematic role in Khader et al. (2003) but were constructed using a pilot questionnaire in which partici pants filled noun or verb gaps in triads with subject object verb relationships, where two related words were provided. The stimuli used in Khader et al. (2003) had more explicit relationships between primes and targets than the
20 stimuli in Balota and Pa ul (1996), and were also verified by native speaker responses in the pilot questionnaire, so that the participants in the priming task were more likely to recognize relationships within trials and also had a greater potential of recognizing patterns of rel ationship across trials. Using these nounnounverb stimuli, a long SOA, and a task that demanded conscious semantic processing and explicit attention to the prime target relationships, Khader et al. (2003) reported an additive priming effect that reduced reaction times significantly. An additive priming effect of 145ms was observed for two nouns on a related verb target (i.e. cheese mous e eat ), with a priming effect of 154ms, compared to a single related noun in the second position (i.e. woodmouse saws ) or in the first position (i.e. cheesecarpenter eat ), which had priming effects of 9ms and 6ms, respectively. Khader et al. (2003) observed a similar pattern of priming effects for verb and noun primes on related noun targets (verbnoun noun presentation i.e. cuts carpenter wood). The much larger additive effects seen in relatedness judgment tasks compared to the LDT experiments in Balota and Paul (1996) with longer SOAs, suggest attentional processing benefits much more from exposure to multiple prim es than does automatic processing (Balota & Paul 1996; Khader et al. 2003). The manipulation of relatedness between unambiguous and ambiguous targets in Balota and Paul (1996) also suggest that longer processing time and increased conscious attention t o the prime target relationships produce greater priming effects when the relationship between the primes and the targets is consciously available to participants. Because the relationships between multiple primes and targets in Balota and Pauls (1996) a mbiguous condition did not converge on single conceptual representations, the prime target relationship was less likely to be apparent to participants completing the experiment. In contrast, the multiple primes converging on a single conceptual representation in Balota and Pauls (1996) unambiguous condition, and primes
21 filling a situational context with targets based on native speaker responses in Khader et al. (2003) would have made the primetarget relationships more likely to be recognized by participa nts, increasing the potential for strategic processing such as expectancy generation to contribute to the priming effect. Presenting multiple primes extends the SOA of trials and the information available to participants, so that expectancy generation i s a potential factor in these results. Expectancy generation refers to a proposed attentional factor in which participants may anticipate specific targets based on the context established by the primes (McRae et al., 2005). Longer SOAs as well as two primes explicitly presented before each target may increase the likelihood of task specific strategies such as expectancy generation, because participants could begin to recognize and anticipate patterns of relatedness between primes and targets (Moss et al., 1995). Expectancy generation is a possible explanation for the much larger additive pri ming effects seen in experiments in Balota & Paul (1996) and Khader et al. (2003) using relatedness judgment tasks and multiple p rimes, but not in all such experiments Where the potential for expectancy generation may be related to differences in the type of relationship between primes and targets, task specific strategic processing, a type of conscious attentional processing, may be a significant factor in the findin gs. Continuous List Priming Whether single or multiple primes are used, the pairing of primes and targets, as well as the identity of targets, are explicit in standard LDT. Though targets and relationships of experimental interest are disguised by nonword trials, and often by additional filler trials, there is a potential for attentional processing to act as an unintended factor in semantic priming effects using this paradigm (McNamara & Altarriba, 1988). To reduce the possibility of attentional processi ng, a variation on LDT has been used in which primes and targets are embedded in a
22 continuous list, with participants asked to respond to every item so that the identity of primes and targets is disguised. This type of task design will be referred to in t his paper as continuous list lexical decision priming, or continuous LDT. Continuous LDT adds to the time between presentation of primes and the onset of targets, as well as introducing variability in the overall SOA of trials, since the response latency to individual primes as well as the planned time between each response and the following list item (response to stimuli interval [RSI]) yield the total SOA. In contrast, standard LDT can be used to examine semantic priming effects at SOAs shorter than RT s to the primes, which have typically shown stronger priming effects of automatic processing than longer SOAs (Neely, 1991). Therefore, standard and continuous LDT offer slightly different insights into automatic factors in language processing. Several s tudies have directly compared standard and continuous LDT using the same stimuli. The average speed of reactions and the size of semantic priming effects are not consistently greater or less in continuous than in standard presentations, yet interesting pa tterns emerge with manipulations of relatedness and SOA. McRae & Boisvert (1998, experiment 1) compared standard and continuous LDT priming for concrete noun pairs in natural (i.e. goose turkey) and artifact (i.e. crayonpencil ) categories, finding signif icant priming effects in both lexical decision conditions, but slower RTs and greater priming in the continuous condition (47ms vs. 29ms for standard LDT). McRae & Boisvert (1998) could not infer a mechanism for this difference, but suggested both attenti onal processing because of the longer SOAs in the continuous task, and stronger activation of primes because of the required response. Perea & Rosa (2002) considered the time course of the effec ts of automatic semantic processing using the same related (s ynonyms, antonyms, and category coordinates) and associated stimuli with
23 standard LDT using 5 different SOAs (66, 83, 100, 116, and 166ms) and continuous LDT using 2 different RSIs (200 and 1,750ms). Standard LDT showed significant priming effects of relatedness at all SOAs except 66ms, where only associative relatedness showed significant priming, while in continuous LDT statistically significant effects of both associa tion and relatedness occurred using a 200ms RSI where targets were presented very quickly following responses to primes but only associated pairs had a priming effect at a longer 1,750 RSI. The findings in Perea and Rosa (2002) suggest that automatic processing can produce semantic priming using both standard and continuous LDT, although the processing time allowed by the SOA or RSI in experimental designs is a factor, with semantic priming absent at a very short SOA and at a very long RSI in a continuous task. McNamara & Altarriba (1988, experiments 1 and 2) used the same group of mediat ed (i.e. lion stripes), directly related (i.e. lion tiger ), and unrelated pairs in separate standard and continuous LDT experiments with a 100ms SOA and RSI, respectively. Equal numbers of participants saw either mediated only or both mediated and directl y related pairs, to compare the priming effect of mediated pairs in the presence of directly related pairs (McNamara & Altarriba, 1988). A priming effect for mediated pairs was observed for the mediated only groups in both experiments, but only in the continuous LDT experiment for the combined mediated and direct groups (McNamara & Altarriba 1988). McNamara & Altarriba (1988) attributed the difference to the elimination of a post retrieval, task specific strategy described as relatedness checking in the continuous LDT task. The Present Study In this study, priming of noun (prime) noun (prime) verb (target) triads was evaluated, similar to Khader et al. (2003). However, this study differs from Khader et al. (2003) in that a continuous list LDT paradigm was used, as opposed to explicitly presented triads with an explicit relatedness judgment task and related stimuli were all agentpatient verb triads as opposed to
24 less restricted contextually related nounnounverb triads used in Khader et al (2003). In this task participants responded to each word, and the time between responses and the following words (RSI) was equal throughout, so that the triads of interest as well as the identify of primes and targets were unk nown to the participants. (See F igure 11 ). Research Questions RQ1. Will activation of paired agent and patient nouns (e.g. car, tourist see Table 1 1) (RR condition) within a familiar situation schema increase the speed of activation for related verbs (e.g. renting) compared to unrelated nouns (UU condition)? RQ2. Will the priming effect of the paired primes (RR) be greater (additive priming) or less (interference) than the effect of the second related primes alone preceded by unrelated nouns (e.g. box, tourist, renting) (UR condition)? RQ3. Will the priming effect of the paired primes (RR) be greater or less than the effect of the first related primes alone followed by unrelated nouns (e.g. car, dentist, renting) (RU condition)? Predictions Based on previous findings (Khader et al., 2003), a priming effect was expected for both participants and items for related agent and patient nouns compared to two unrelated noun primes preceding the same target verbs, using a continuous LDT task (RQ1). If observed, this will demonstrate a priming effect on verbs of multiple thematic role nouns related via familiar situation schemas. No specific predictions were made for RQ2 or RQ3, since previous findings indicating an additive effect for two related primes (RR) compared to single related primes were th e result of different priming paradigms. However, if the RR condition produced RTs greater than single primes in both the first and second position paired with unrelated noun primes (UR and RU) (both RQ2 and RQ3), then it would indicate an additive priming effect for the RR
25 condition. Previous findings regarding multiple primes strongly suggest that priming effects of single related primes are likely to be greater in the second than in the first position (where the unrelated prime appears between the related prime and the target), so that an additive priming effect can only be inferred if the priming effect in the RR condition is significant greater than both the UR (RQ2) and RU (RQ3) conditions (Balota & Paul, 1996; Khader et al., 2003). See T able 1 2 fo r a list of planned post hoc mean comparisons used to test the three research questions.
26 Table 1 1. Sample stimuli and experimental conditions Target verbs were presented in present progressive ( ing) form in each experimental condition, preceded by t wo primes that were either related or unrelated, creating four relatedness conditions. Related primes fill the thematic roles of agent and patient in common situation schemas around the verbs. The comple te stimuli list is included in A ppendix A. Table 1 2. Planned mean comparisons Research Question Mean RT for Condition: Vs. Mean RT for Condition: RQ1 Related+related primes [RR] Unrelated unrelated primes [UU] RQ2 Related+related primes [RR] Unrelated+related primes [UR] RQ3 Related+related primes [RR] Related+unrelated primes [RU] These three post hoc mean comparisons were planned to test the significance of differences between the mean reaction times (RTs) of experimental conditions. Results of these comparisons were used to test pr edictions related to the three research questions (RQs). Experimental Condition Prime 1 Prime 2 Target Related+related primes [RR] car tourist renting Unrelated+r elated primes [UR] box tourist renting Related+unrelated primes [RU] car dentist renting Unrelated+unrrelated primes [UU] box dentist renting
27 Figure 1 1. Continuous list lexical decision task schematic. Participants were instructed to respond to every letter string in the priming experiment (1000 items total) by pressing keys marked Y for yes, for real English words, or N for no, for pronounceable nonwords. Experimental triads (i.e. thiefmoney steal ) were embedded in the list along with filler words (i.e. blanket) and nonwords (i.e. falping) so that items of interest were disgui sed from participants.
28 CHAPTER 2 METHOD Participants Population and Criteria for Inclusion To judge the relatedness of stimuli, 15 young adults were recruited to complete a survey using a rating scale. All subjects recruited for the relatedness survey completed the task. To complete the priming experiment, 47 participants were recruited. All recruited participants completed informed consent procedures as approved by the University of Florida Institutional Review Board (IRB). All participants who comp leted the relatedness judgment survey or the priming experiment were young adult, right handed, monolingual English speakers, age 1830, with no history of learning disabilities or neurological disorders. Two participants were found not to meet the study criteria upon review and did not complete the priming experiment. One was left handed. One participant was bilingual, with simultaneous acquisition in early childhood and lifelong use of both Spanish and English. Recruitment of Participants Participant s were recruited from the LIN SLHS Participant Pool at the University of Florida, whose listings are posted online at http://slhs.phhp.ufl.edu/labs/language brain/pool/. In addition, fliers were posted at the Speech and Hearing Clinic in Dauer Hall at the University of Florida. Participants received compensation in the form of either course credit from their instructors, based on their syllabus, or a $5 gift card. Stimuli Stimuli Development Stimuli consisted of prime prime target triads in which all primes were nouns and all targets were verbs in the present progressive ( ing) form. Relatedness of the primes to their targets was
29 manipulated and in the related condition one prime was a typical agent, and the other a typical patient, of the target verb. Primes and targets were related via familiar situation schemas in which they commonly occur together. Target verbs were also paired with unrelated primes. The effect of animacy was controlled by matching related and unrelated primes for animacy on a one to one basis. Related and unrelated primes were matched as closely as possible on the psycholinguistic variables of written frequency, length (in letters, phonemes, and syllables), concreteness, printed familiarity, imageability and age of acquisition rat ings. Psycholinguistic data as well as unrelated nouns were collected using the MRC Psycholinguistic Database (Clark, of present day American English. This text provides the frequency of occurrence of words in a 1million word selection of text from a variety of sources, now known as the Brown Corpus (after Brown University, where it was produced). Printed familiarity (FAM), concreteness (CNC), and imageability (IMG) values were derived by merging normative data from three sources, the Paivio Norms (an unpublished expansion of Paivio Yuille & Madigan, 1968), Toglia and Battig (1978), and Gilhooly and Logie (1980), by a method described by Coltheart (1981). Age of ac quisition (AOA) values were drawn from Gilhooly & Logies (1980) normative data, mathematically transformed to match the 100 700 point range used for the other rated variables. Average values for these variables were collected for both related and unrelat ed primes, as well as the verb targets and the verbs in their infinitive forms, for which ratings data are more often available. There were two stages in stimuli development; initial development and selection of potential stimuli triads, and a commonality rating survey used to measure relatedness judgments.
30 Stimuli Questionnaire Potential related stimuli were drawn from several sources. Most (31) were developed in a thematic role completion questionnaire, in which participants responded to a series of r elated questions and statements asking for single word responses fitting a thematic role. The thematic role questionnaire was completed by 30 participants drawn from the undergraduate population of the University of Florida. Participants were offered com pensation in the form of course credit from their instructors according to their course syllabus. Participants completed a University of Florida IRB approved consent form before engaging in the task. The questionnaire was presented on standard computer m onitors using Microsoft Office Access (2007). Questions consisted of a series of related questions and statements asking participants to enter single word responses fitting one of five thematic roles; verb, agent, patient, instrument, or location, with one to five response spaces each. At least one thematic role filler was provided for each series of questions, and participants provided related fillers for two or more additional roles. Figure 2 1 shows a sample series of questions from the practice section of the stimuli development questionnaire, with example respons es. Verbs were either provided or contextually guided to fit the intended thematic role structure. All verbs allowed transitive syntactic argument structures (NP V NP) and had twoplace minimum thematic grids (agent, patient), which refers to the thematic roles that must be filled by syntactic arguments in a sentence (Levin, 1993; Shapiro et al., 1991). Verbs that were provided were drawn from Levins (1993) English verb classes and alt ernations and from VerbNet v.3.1, accessed on the Unified Verb Index Web site (Kipper, Korhonen, Ryant, & Palmer, 2008; Palmer & Kipper 2010). The thematic role questionnaire yielded a large number of potential agent patient verb triads (more than 5,000 unique triads). Potential stimuli for the relatedness judgment survey were chosen from among the most frequently occurring responses, in addition to meeting
31 psycholinguistic criteria described below. An additional seven related triads were created by co mbining semantically related agent verb and patient verb dyads identified in earlier research (Edmonds & Mizrahi, submitted; McRae et al., 2005). Attempts were made to avoid the inclusion of strongly class ambiguous words, because the priming experiment p rocedure used in the present study does not provide contextual cues to bias ambiguous words for either noun or verb meanings, and because processing of class ambiguous words has been shown to be different than for unambiguous words (Federmeier, Segal, Lomb rozo & Kutas 2000). Because English verbs frequently have corresponding noun forms and vice versa (i.e. to butcher, a butcher ), the dominant usage of potentially am biguous words Frequency analysis of English usage which gives frequency counts of English words by part of speech (e.g. noun, verb, adjective, etc.). All words used as nouns in this study were more frequent than verb forms by a ratio of 6 or better. All words used as verbs, which are as a class used less frequently than nouns, were dominant over noun forms by a ratio of 2 or better, with one conflict disregarded because it is not consistent with modern usage ( healing, produced with equal frequency as a noun and a verb in Verb forms were typically much higher in frequency than nouns when all morphological forms were collapsed by word class, since English verbs are inflected in more ways than nouns (e.g., fixing had a verb:noun ratio of 4.5, but at the lemma level, all verb forms including fix, fixing, fixed, etc. had a ratio of 36 to all noun forms including fix, fixing) Association norms were not used to test the relatedness of stimuli items by design. Specific semantic and event knowledge were targeted rather than direct associations, which were minimized to avoid unintended effects of association. Because it is possible for thematically -
32 related words to also be strongly associated, p otential stimuli triads were checked for forward association in prime prime and prime target pairings (three pairings per triad) where available by using the University of South Florida free association norms which were collected using a written task where participants r ead each cue word and filled in a blank with the first word that came to mind ( Nelson, McEvoy, & Schreiber, 1998) Item pairings with association ratios greater than .10, meaning that 10% of respondents to a cue named the target, were avoided for this stu dy. For verbs, both infinitive and present progressive ( ing) forms were considered for the purpose of judging association, although verbs were presented only in the present progressive form as targets in the priming study. This was a conservative approa ch intended to prevent phonological activation through strongly associated class ambiguous items. For example, cook is ambiguous for noun and verb, and has a strong association from chef via its synonymous noun meaning, so that the pairing of chef with the verb form cooking was excluded even though chef and cooking are not strongly associated. Stimuli were also drawn from those reported in Ferretti, McRae, & Hatherell (2001), for which verb agent and verbpatient pairs were checked for associ ation. All three forward pairings were not available in these existing published association studies for all potential stimuli triads. Stimuli were not necessarily rejected for this reason if there was intuitive reason to believe that a strong associatio n was unlikely (e.g. landlordevicting) but the percentage of pairings that did not have documented association norms was recorded. Association norms were available for 83% of the forward pairings. One direct pairing exceeded a 0.10 association ratio ( boss employee, 0.147). Two prime target pairings (in which an intervening prime was used in this experiment) exceeded a .1 association ratio in either the infinitive or present progressive verb form ( dicerolling, 0.124, and thiefsteal 0.388). All other forward pairings had a lower than 0.10 association ratio. Backward
33 associations from prime 2 to prime 1 were also recorded when available, and norms were available for 77% of pairings. Five backward prime pairings had association ratios greater than 0.10. Relatedness Survey Potential stimuli triads created in the stimuli questionnaire phase were tested for semantic relatedness using a rating survey. The goal of this survey was to guide final selection of stimuli for the priming experiment and to demonstrate that the selected role fillers accurately match participant judgments about the commonality of the scenarios formed by the triads. Related agents and patients and their psycholinguistically matched unrelated nouns were mixed in four relatedness co nditions. It was expected that the related agent related patient [AP] condition would be rated significantly more related than the related agent, unrelated patient [AU], unrelated agent, related patient [UP], and unrelated agent, unrelated patient [UU] conditions; that the UU condition would be rated least related, and that they AU and UP conditions would be rated significantly less related than the AP condition and significantly more related than the UU condition. It was further expected that the AU and UP conditions would not be significantly different in their average relatedness ratings. Similarity of the mixed condition AU and UP triads was desirable because the presentation order of primes in the priming experiment was balanced across agent patienttarget and patient agent target to further obscure the identity of experimental items within the lexical decision task, and to reduce the likelihood of syntactic based expectancy generation. Participants. The relatedness survey was completed by 15 native English speaking participants. The participants ranged in age from 19 to 25, with an average age of 21. Participants were recruited from the Gainesville, FL and University of Florida communities and completed a University of Florida IRB approved consent form before completing the survey.
34 Materials. The survey included 152 triads developed using 38 verbs. The 38 related (AP) triads were identified first from the stimuli development questionnaire phase described above, then each related agent or patient noun was matched with an unrelated noun. The related and unrelated nouns were combined with the verbs in each of the four conditions, yielding the total of 152 triads. Unrelated items were matched individually to agent and patient nouns based on animacy and on frequency range (low: 1 10; medium: 1160; high >60 after Merritt, DeLosh & McDaniel, 2006), and as closely as possible on length in letters, phonemes and syllables. Related and unrelated agents and patients were matched by group across the psycholinguistic variables of age of acquisition, frequency, concreteness, familiarity, and imageability as described above for stimuli development. Data for verbs in the present progressive ( ing) form (e.g. wearing) used in this study are frequently unavailable on these variables; therefore, data for the infinitive (basic) form (e.g. wear ) were collected as well. From the 38 sets rated, a final list of 30 sets was selected for the priming experiment. From the pool of potential related triads, stimuli were s elected based on two additional criteria. Where potential verbs were semantically related (i.e., cut, slice ) or nouns occurred in more than one potential triad, the option that best fit the other selection criteria was retained. Procedure. Participants rated the triads on a survey form. Triads were presented without syntactic context, but always in agent verb patient order (i.e. woman wearing blouse ). Verbs were presented in the present progressive ( ing) form, since that is how they would be r epresented in the priming portion of the study. Ratings were obtained for each verb in all four relatedness conditions; AP, AU, UP, and UU. The list was divided into four sections such that each verb occurred only once in each section, and an equal numbe r of triads in each of the four conditions occurred in each section. The order of presentation within each set was randomized,
35 and four lists were created with the presentation order of the sections changed. The choice of list for each participant was ch osen by recruitment order. Participants were asked to rate how common it is for the scenarios described by each triad to occur. Similar to previous investigations (Edmonds & Mizrahi, submitted; Ferretti et al., 2001; McRae, Ferretti, & Amyote, 1997), tri ads were rated by circling a number on a 7 point ordinal scale with 1=not common at all, and 7=highly common. The average relatedness rating of triads across four conditions was rated for 38 sets sharing a verb, totaling 152 triads. Results. Table 2 1 shows the mean rating and standard deviation for each condition. Complete relatedness ratings data for all exper imental triads are included in A ppendix B. 2(5) = 26.81, p < .001) indicating that the assumption of sphericity in the condition variances was not met, as expected, and adjustment to degrees of freedom was required (Field, 2009). Repeatedmeasures ANOVA was performed with a HuynhFeldt correction, which is a conservative procedure to adjust for the unequal variance between conditions (Huynh Feldt Epsilon = .716) (Field, 2009). This measure showed a significant effect of relatedness (F(2.15,62.32) = 491.98, p < .001) across conditions. A Sidak post hoc analysis with Holms sequential correction (Holms, 1979) revealed significant differences between the AP condition and all other conditions ( p < .001 for all comparisons) and the UU and all other conditions ( p < .001 for all comparisons). Holms (1979) correction is a method of adjusting significance values for multiple related tests As predicted there was no significant difference between the AU and UP conditions ( p = .877). See Table 22 for details. Sti muli List Table 2 3 shows the average length of all stimuli in this priming experiment. Due to the nature of the continuous list LDT priming paradigm used in this study, participants responded to
36 three forms of stimuli. In addition to experimental stimuli which consisted of the primes and targets developed in the stimuli development phase, filler words and nonwords were included Experimental s timuli From the results of the relatedness survey, a final list of 30 verbs with related and unrelated primes was selected for the priming experiment. The experiment al stimuli list is included in A ppendix A. Mean values on psycholinguistic variables are presented in T able 2 4 by stimuli type. To verify that related and unrelated primes were not significantly diff erent on any of these psycholinguistic variables, a series of one way ANOVA comparisons of related and unrelated prime 1s and related and unrelated prime 2s was performed. These analyses did not find a difference between these stimuli groups on any of the psycholinguistic factors at a 0.05 significance level (see T able 2 5). These results verify that the related and unrelated primes were similar on psycholinguistic factors that are potential factors in priming but were not of interest in this study. A c omplete list of psycholinguistic data for all exper imental stimuli is included in A ppendix C. Filler w ords Non experimental filler words were included in the task to separate the experimental triads, reduce the likelihood of recognizing the pattern of rela tionships in experimental triads, and to keep the related unrelated proportion low. The relatedness proportion is the ratio of the trials in which targets are preceded by a related word to the total number of trials (Ferretti et al. 2001; McRae et al ., 2005; Neely, 1991). A low relatedness proportion reduces the likelihood that participants will recognize or anticipate the pattern of relatedness used in the experimental trials. A total of 140 filler words were drawn from the MRC Psycholinguistic Database (Clark, 1997), similar to the process used to select unrelated primes. The relatedness proportion cannot be calculated in the same way for continuous LDT as it is for standard LDT because primes are
37 not obviously paired with targets (McRae & Boisvert, 1998). However, an attempt was made to estimate the relatedness proportion by calculating the ratio of trials that were deliberately preceded by a related word to the total number of trials, yielding an estimated relatedness proportion of 0.18, which can be considered low for comparison to studies examining similar thematic role relationships (McRae et al., 2005). Nonwords Nonwords consisted of letter strings that follow the orthographic rules of English but do not exist in the lexicon. Pronounceable non words (i.e. pess ) were chosen over the pseudohomophone type (i.e. brane ) because they result in easier discrimination in an LDT task, which supports the goal of minimizing overall trial duration in this experimental design, and because they have been found to increase semantic priming effects when compared to a pseudohomophone condition (Yap, Tse & Balota, 2009). Nonwords were individually matched to words for length in letters and syllables to avoid unintended effects of length between word and nonword i tems during the experimental task Phonemes are properties of the verbal representations of words, not their orthographic forms, so length in phonemes is not applicable for printed nonwords. Nonwords were either drawn from the ARC nonword database (Rastl e, Harrington & Coltheart, 2002) or by changing one or more lett ers in a real word (e.g. g n syllable words, and were used to form multisyllable words by combining or by adding real English morphemes present in the word list ( e.g. ing, ist ). Apparatus Participants were seated at a standard desk before a Dell Latitude D820 laptop with a 15 inch monitor. Two keys were marked for responses on the lexical decision task; the < key was
38 labelled with a Y for Yes responses, an d the > key with an N for No responses. This experiment was conducted using DirectRT v2006 software (Empirisoft). Procedures Continuous list lexical decision task. Experimental triads were embedded in a presentation list in random order. Participa nts were instructed to read and respond to every letter string presented on a screen by pressing a yes or no key to indicate whether it was a real English word. Participants were given no information about which items were of experimental interest. L etter strings remained on the screen until a response was given, followed by a response stimulus interval [RSI] of 100ms consisting of a blank screen before the next item in the list appeared. See F igure 1 1 for a schematic of the task presentation. Parti cipants completed an instruction section with an experimenter, including four sample letter strings. The sample items provided feedback indicating correct and incorrect responses, and repeated sample items if the incorrect Y or N key was pressed. Aft er they conveyed understanding of the correct responses for words and nonwords, participants completed a practice section of 25 letter strings. During the practice section, items advanced automatically following correct responses, and feedback was given following incorrect responses. Error feedback consisted of either Error. That was a real English word, and the correct answer is Y or Error. That was NOT a real English word, and the correct answer is N., followed by Press enter to continue practic e items.. During the practice section participants were given feedback if they responded faster than 100ms (Please wait for the letter string!) or slower than 1200ms (Please try to respond faster!), with a delay of 2000ms before the next practice i tem appeared. Following the practice section participants were given an opportunity to ask questions about the task and conveyed understanding before starting the experiment. Participants were informed that they would no longer receive any feedback befor e they began the experiment. A text version of
39 a ll instructions in included in A ppendix D. All experimenters received training and care was taken to disguise the true interests of the experiment. Experimenters addressed any questions about the task, but any questions about the experiment were deferred to a debriefing session at the conclusion of the visit. Questions about the design and predictions of the experiment were addressed only with regard to word recognition in the word/nonword context. Partic ipants were probed about whether they had noticed anything of interest about the list items, and no participant indicated recognition of the thematic role relationships in the experimental triads. During the experiment participants saw 1000 letter strings consisting of 120 experimental triads (360 word presentations), 140 filler words, and 500 nonwords. The ratio of words to nonwords was 0.5, such that yes and no responses were equally likely for each letter string. The experimental triads were divid ed into four lists such that each target verb appeared only once in each list and the experimental conditions were balanced across the lists. The order of presentation of the lists, and of the experimental and nonexperimental letter strings within each list, were randomized for each participant so that any effects of learning or fatigue were distributed across all targets and conditions. The lists were organized with a semi random number of letter strings appearing between each of the experimental triads, with a minimum of 3 and an average of 5.3 nonexperimental letter strings between each triad. To reduce fatigue, participants were given two 5 minute breaks, and the experiment was completed in less than 30 minutes. Letter strings appeared in the center of the monitor screen in Times New Roman, size 32 black font, with a white background. An asterisk appeared for 500ms as a warning signal before the practice section, before the first letter string at the beginning of the experimental task and after each break. Each letter string remained in the center of the screen until the participant
40 entered a response. Responses were followed by a blank screen for 100ms (RSI) before the next letter string in the list appeared. Design and Analyses A repeated measures design was used, with four relatedness conditions. Each participant completed all conditions, distributed throughout the experimental task, so that participants served as their own controls. Participant results were initially analyzed for overall accuracy of lexical decisions. A minimum of 95% accuracy was required for inclusion in further analysis. Lexical decisions for both primes and the target in each experimental triad were examined, and RTs to targets were excluded from further analysis for any tr iad that had one or more lexical decision errors. It is standard to exclude trials containing decision errors from further analysis (Ferretti et al., 2001; Jones, 2010; Khader et al., 2003; Perea & Rosa, 2002; Vigliocco et al., 2003). In standard LDT tas ks a measurable behavioral response is elicited only for target items. In continuous list LDT tasks with single primes, trials including errors on either the prime or target have been excluded from RT analysis (McNamara & Altarriba, 1988; Moss et al., 1995). This study used a similarly conservative approach intended to reduce potential confounding effects from failure to attend to or correctly process any primes. Participant Analysis Decision latencies. Response latencies (RTs) that fell greater than tw o standard deviations above or below the mean for each participant by condition were excluded from further analysis. Effects of relatedness were examined using repeated measures ANOVA. Then, Tukey LSD post hoc analyses with a Holms (1979) correction we re performed to evaluate potential differences that corresponded to our research questions (RR v UU [RQ1], RR v UR [RQ 2] and RR v RU [RQ3]; refer to T able 1 2).
41 Error rates. The mean error rate for each participant by condition was determined. Mean err or rates were compared for effects of relatedness using repeated measures ANOVA. Based on the results of similar research, no significant effect of relatedness on error rates was anticipated (Ferretti et al., 2001; Hare et al., 2009; Jones, 2010). Item An alysis. Decision latencies. Response latencies (RTs) that fell greater than two standard deviations above or below the mean for each target by condition were excluded from further analysis. Effects of relatedness were examined using repeated measures AN OVA. Then, Tukey LSD post hoc analyses with a Holms (1979) correction were performed to evaluate potential differences that corresponded to our research questions (RR v UU [RQ1], RR v UR [RQ 2] and RR v RU [RQ3]; refer to T able 1 2). Error rates. The mean error rate for each item by condition was determined Mean error rates were compared for effects of relatedness using repeated measures ANOVA. Based on the results of similar research, no significant effect of relatedness on error rates was anticip ated (Ferretti et al., 2001; Hare et al., 2009; Jones, 2010).
42 Table 2 1. Mean ratings, standard deviations (SD) and ranges on a relatedness judgment survey by condition (N=15) Condition Example Mean SD Range AP TOURIST RENTING CAR 6.6633 .20053 6 .2667 6.9333 AU TOURIST RENTING BOX 2.5635 .87987 1.4000 4.9333 UP DENTIST RENTING CAR 2.3508 .71863 1.4000 3.6667 UU DENTIST RENTING BOX 1.6419 .39848 1.1429 2.4667 AP=related agent related patient condition; AU=related agent unrelated pat ient condition; UP=unrelated agent, related patient condition; UU=unrelated agent unrelated patient condition. N=number of participants. Participants rated each triad on an 7 point ordinal scale with 1=Not common at all and 7=highly common. Table 2 2. M ean comparisons of relatedness judgment ratings by condition Conditions compared Mean Difference SE Sign. AP AU* 4.100 .162 .000 AP UP* 4.313 .137 .000 AP UU* 5.021 .078 .000 AU UP .213 .199 .877 AU UU* .922 .144 .000 UP UU* .709 .120 .00 0 *Significant mean difference at the .05 level. AP=related agent related patient condition; AU=related agent unrelated patient condition; UP=unrelated agent, related patient condition; UU=unrelated agent unrelated patient condition Sidak post hoc analyses of mean differences in relatedness judgment ratings were performed to verify that the stimuli triads were significantly different in relatedness. As expected, all differences were significant except the AU UP conditions, which were combined in the primi ng experiment so that agents and patients appeared equally often in prime 1 and prime 2 positions. Table 2 3. Length of letter strings in priming experiment Thematic Role NLET NPHN NSYL N Targets (verbs) 8.03 6.43 2.33 30 Related Prime 1 5.63 4.67 1.73 30 Unrelated Prime 1 5.63 4.57 1.73 30 Related Prime 2 7.17 5.83 2.30 30 Unrelated Prime 2 7.20 5.80 2.27 30 All Experimental 6.95 5.62 2.12 360 Filler Words 6.57 5.21 2.03 140 All Words 6.84 5.51 2.09 500 Nonwords 6.82 n/a 2.06 500 Mean length in number of letters (NLET), phonemes (NPHN) and syllables (NSYL) of all items in the priming experiment list. NPHN was not applicable to nonwords. The total number (N) of all experimental words included all primes and targets. Each target appeared four t imes in the list, and each prime appeared twice; with the addition of filler words, there were an equal number of words and nonwords in the list.
43 Table 2 4. Group means for controlled psycholinguistic variables Priming Role AOA N KFFRQ N CNC N FAM N IMG N Targets (Verbs) 306.80 5 20.52 29 451.00 3 532.80 5 472.60 5 Infinitive form 365.00 6 29.03 29 443.20 15 530.69 16 484.81 16 Related Prime 1 306.09 11 48.82 28 553.37 19 541.05 20 568.00 19 Unrelated Prime 1 390.22 9 40.00 30 550.92 25 511.56 25 542 .64 25 Related Prime 2 374.40 10 58.67 27 529.88 17 531.06 18 541.00 17 Unrelated Prime 2 347.00 11 46.07 30 532.89 18 537.58 19 545.28 18 Related and unrelated primes were matched as closely as possible on psycholinguistic variables known to influence priming. Data were collected wherever possible (N = the number of data points included in each mean) on age of acquisition (AOA), frequency (KFFRQ), concreteness (CNC), familiarity (FAM), and imageability (IMG). Data on verbs in the present progressive ( i ng) and the more frequently available infinitive forms provided for informational purposes. Table 2 5. Statistical comparisons for controlled psycholinguistic variables. Related v. unrelated prime 1 Related v. unrelated prime 2 Factor F(1) sig. ( p = ) F(1) sig. ( p = ) AOA 2.24 0.152 0.52 0.481 KFFRQ 0.32 0.574 0.34 0.563 CNC 0.01 0.918 0.03 0.872 FAM 3.37 0.073 0.13 0.724 IMG 1.25 0.269 0.05 0.823 NLET 0.00 1.000 0.01 0.938 NPHN 0.04 0.844 0.00 0.949 NSYL 0.00 1.000 0.02 0.892 *N o comparisons have a significant mean difference at the .05 level AOA=age of acquisition; KFFRQ=frequency; CNC=concreteness; FAM=familiarity; IMG=imageability; NLET=number of letters; NPHN=number of phonemes; NSYL=number of syllables; sig.=significance. A series of one way ANOVAs compared psycholinguistic variable means for matched related and unrelated primes. As expected, no comparisons are significant, verifying that these variables known to influence priming are controlled across conditions.
44 Figure 21. Stimuli development questionnaire. This figure presents a sample series of questions and examples of responses from the stimuli development questionnaire. Participants were instructed to provide single word responses in the boxes but were not othe rwise constrained. Some thematic role fillers were provided (e.g. hammer), while some words within in a series of questions were filled in by the participants own responses (e.g. hang) so that participants provided multiple responses within each situatio n schema.
45 CHAPTER 3 RESULTS The 45 participants who completed the study had a mean age of 20.75 (SD 2.44) years, and 20% were male. The requirement of 95% accuracy on the lexical decision task was not met by 5 participants, and their data were excluded f rom further analysis. The remaining 40 participants had a mean age of 20.55 (SD 2.09) years, and 18% were male. The overall accuracy of lexical decisions was 98%, which falls in a range typical of lexical decision tasks (Ferretti et al., 2001; Jones, 2010; McRae & Boisvert, 1998). An error on lexical decisions for either prime or the target occurred on 6% of experimental triads. For the data analyzed after omitting errors, participants responded f aster to words than to nonwords on 39,138 correct respons es (618ms v. 665ms). The average SOA from presentation of prime 1 to onset of the target for experimental triads was 1432ms, consisting of the mean RT to primes (616.15ms) plus the 100ms RSI, for two primes per triad. Participant Analysis Decision latenci es. Target responses that were greater than 2 SDs above or below the condition mean for each participant were excluded from participant analysis, consisting of 6% of data. In combination with responses excluded due to lexical decision errors in a triad ( not a sum because some responses qualified for exclusion for both reasons), 10% of total data were excluded from participant analysis. An unplanned series of one way ANOVA analyses comparing the condition means (as used in further analysis, with trials that have any lexical decision error excluded) with the condition means when only errors to the targets are excluded found no significant difference for any of the four experimental conditions ( p>.05 for all comparisons).
46 A repeated measures ANOVA showed a s ignificant effect of relatedness across participants (F(3 ,39) = 4.62, p 2(5) = 8.34, p = .139) indicating that the assumption of sphericity in the condition variances was met and no adjustment to degrees of freedom was required (Field, 2009). Participants had a mean RT of 557.57ms (SD 66.94) to targets in the RR condition, a mean of 560.70ms (SD 70.81) to targets in the UR condition, a mean of 573.84ms (SD 82.60) to targets in the RU condition, and a mean of 574.24ms (SD 80.37) to ta rgets in the UU condition (see Table 31 and F igure 3 1). T here was a difference of 16.73ms between RR and UU means for participants (RQ1). A Tukey LSD post hoc analysis of the RR UU mean comparison showed a significant difference ( p = .021) using a significance level of .025 based on Holms (1979) a djustment for multiple comparisons. Participants showed a mean difference of 3.20ms between RR and UR means (RQ2). This difference did not meet the p = .05 level required for significance in this study, in a Tukey LSD post hoc analysis of the RR v. UR m ean comparison ( p = .570). Participants had a mean difference of 16.34ms between RR and RU means (RQ3). A Tukey LSD post hoc analysis of the RR RU mean comparison showed a significant difference ( p = .012) using a significance level of .0125. Figure 3 1 shows a graph of these mean differences. Table 3 2 summarizes the RT mean differences for participants for each of the three research questions posed in this study, actual pvalues in post hoc analyses, and required pvalues to achieve significance based on the adjustment for multiple comparisons used in this study. Error rates. The mean error rate for each participant by condition was calculated. The results of a repeated measures ANOVA showed that the effects of relatedness on error rates was not sig nificant for participants (F(3 ,39 ) = 1.64, p = .183). This result was expected based on
47 findings in previous research as well as the low overall error rate of 2% ( Ferretti et al., 2001; Hare et al., 2009; Jones, 2010). Item Analysis Decision latencies. T arget responses that were greater than 2 SDs above or below the condition mean for each item and excluded from item analysis, consisting of 5% of data. In combination with responses excluded due to lexical decision errors in a triad (not a sum because som e responses qualified for exclusion for both reasons), 10% of total data were excluded from item analysis. An unplanned series of one way ANOVA analyses comparing the condition means (as used in further analysis, with trials that have any lexical decision error excluded) with the condition means when only errors to the targets are excluded found no significant difference for any of the four experimental conditions ( p>.05 for all comparisons). A Repeated measures ANOVA showed a significant effect of related ness across items (F(3 ,29 ) = 3.67, p 2(5) = 6.39, p = .271) indicating that the assumption of sphericity in the condition variances was met and no adjustment to degrees of freedom was required (Field, 2009). RTs to items had a mean of 555.88ms (SD 32.71) in the RR condition, a mean of 564.61ms (SD 44.45) to targets in the UR condition, a mean of 573.39ms (SD 47.44) to targets in the RU condition, and a mean of 572.75ms (SD 51.17) to ta rgets in the UU conditio n (see Table 3 3 and F igure 3 1). There was a mean difference of 16.86ms between RR and UU means for items (RQ1). A Tukey LSD post hoc analysis of the RR UU mean comparison showed a significant difference ( p = .011) using a significance level of .025 b ased on Holms (1979) adjustment for multiple comparisons. Items showed a mean difference of 8.73ms between RR and UR means (RQ2). This difference did not meet the p = .05 level required for significance in this study, in a Tukey LSD post hoc analysis of the RR v. UR mean comparison ( p = .107). Participants had a mean
48 difference of 17.51ms between RR and RU means (RQ3). A Tukey LSD post hoc analysis of the RR RU mean comparison showed a significant difference ( p = .006) using a significance level of 0125. Figure 31 shows a graph of these mean differences. Table 34 summarizes the RT mean differences for items for each of the three research questions posed in this study, actual pvalues in post hoc analyses, and required pvalues to achieve signific ance based on the adjustment for multiple comparisons used in this study. Error rates. The mean error rate for each item by condition was calculated. The results of a repeated measures ANOVA showed that the effects of relatedness on error rates was not significant for items (F(3 ,29) = 1.43, p = .240). This result was expected based on findings in previous research as well as the low overall error rate of 2% (Ferretti et al., 2001; Hare et al., 2009; Jones, 2010).
49 Table 3 1. Mean reaction times by c ondition for participant results (N=40) Condition Participant mean (ms) SD RR 557.51 66.94 UR 560.70 70.81 RU 573.84 82.60 UU 574.24 80.37 RR=related related primes condition; UR=unrelated related; RU=related unrelated; UU=unrelated unrelated. N=numb er of participants. SD=standard deviation. Table 3 2. Condition comparisons for participant results (N=40) Comparisons Mean Difference Actual p values Required p values to achieve significance* Significance RR v UU (RQ1) 16.73 .021 .025 sig RR v UR ( RQ2) 3.20 .570 .05 ns RR v RU (RQ3) 16.34 .012 .0125 sig *Adjusted for multiple comparisons based on Holm (1979). Post hoc mean comparisons related to each of three research questions (RQs) for participant results. RR=related related primes condition; UR=unrelated related; RU=related unrelated; UU=unrelated unrelated. N=number of participants. Sig=significant difference; ns= nonsignificant difference. Table 3 3. Average reaction times by condition for item results (N=30) Condition Item mean (ms) SD RR 555.88 32.71 UR 564.61 44.45 RU 573.39 47.44 UU 572.75 51.17 RR=related related primes condition; UR=unrelated related; RU=related unrelated; UU=unrelated unrelated. N=number of items. SD=standard deviation. Table 3 4. Condition comparisons for item results (N=30) Comparisons Mean Difference Actual p values Required p values to achieve significance* Significance RR v UU (RQ1) 16.86 .011 .025 sig RR v UR (RQ2) 8.73 .107 .05 ns RR v RU (RQ3) 17.51 .006 .0125 sig *Adjusted for multiple compa risons based on Holm (1979). Post hoc mean comparisons related to each of three research questions (RQs) for item results. RR=related related primes condition; UR=unrelated related; RU=related unrelated; UU=unrelated unrelated. N=number of items. Sig=sig nificant difference; ns=nonsignificant difference.
50 Figure 3 1. Graphs of participant and item means. RR=related related primes condition; UR=unrelated related; RU=related unrelated; UU=unrelated unrelated. Mean reaction times (RTs) in milliseconds (ms) for participants and items in each relatedness condition, with standard deviation indicated by error bars. See Tables 31 and 33. Figure 3 2. Graphs of mean differences for each research question. *=significant difference based on Holms (1979) adjustment for multiple comparisons. RR=related related primes condition; UR=unrelatedrelated; RU=related unrelated; UU=unrelated unrelated. Mean reaction times (RTs) differences in milliseconds (ms) for participants and items related to each of three r esearch questions (RQs). See Tables 33 and 34.
5 1 CHAPTER 4 DISCUSSION The results of this priming experiment were very similar for participant and item analyses, with a few interesting differences. Mean RTs for each condition and mean differences wer e similar for participants (N=40) and items (N=30), with significance found for the same comparisons (RQ1 and RQ3) for both. Standard deviations were lower for items than for participants in all conditions, and there was greater significance (lower p valu es) for items on all mean comparisons addressing the research questions (although RQ2 did not approach significance for either participants or items). This overall pattern of results suggests greater variability among participants than between items. Although few researchers report condition means and SDs for both participants and items (despite the widely accepted practice of reporting both participant and item ANOVAs), RaczaszekLeonardi et al. (2008) found significant variability in participants but no t in items when comparing results across multiple SOAs in a cross modal LDT. R esearch Q uestion 1 Statistically significant mean differences between RR and UU for both participants and items shows that there is a facilitation effect of approximately 17ms fo r two noun primes to related verbs, with the thematic role relationships of agent, patient and action used in this study. This is an expected result because previous studies have observed priming effects of single thematically related noun primes to verbs and of two related primes to targets, when the primes were explicitly associated to the targets (Balota & Paul, 1996; Edmonds & Mizrahi, submitted; Ferretti et al., 2001; Khader et al., 2003; McRae et al., 2005). Related primes and targets that had pre viously produced semantic priming effects using standard LDT have produced similar
52 results, though often with smaller semantic priming effects, using the continuous LDT paradigm used in this study (McNamara & Altarriba, 1988; Perea & Rosa, 2002) R esearch Q uestion 2 RQ2 addressed the difference between the priming effect two related primes (the RR condition), and one related prime preceded by an unrelated noun (the UR condition). The difference in the mean RTs between RR and UR was not significant in this s tudy for either participants or items. In this population of healthy, neurologically typical young adult English speakers, both related primes were not necessary to produce the priming effect observed in RQ1, when compared to single related primes present ed immediately before the targets. There was neither additive priming nor interference created by the addition of a second related noun prime. This condition approximates single prime target pairings in a continuous LDT paradigm, and can be compared to previous findings related to thematic roles using single word priming paradigms (Edmonds & Mizrahi, submitted; McRae et al., 2005). Given the continuous list design, in which the primes were not explicitly grouped with the targets, the UR condition was s imilar to a continuous list, single prime design with the caveat that the nouns preceding the related prime were never random, being concrete nouns fitting the thematic frames of agent or patient, though not related to the situation schema represented by t he related prime and target verb. The priming effect of mixed agent and patient primes on verbs in the UR condition in this study (13ms in the participant analysis, 8ms for items) was slightly smaller than the effect observed by McRae et al. (2005) using a pronunciation task (18ms for agents and 22ms for patients). Edmonds & Mizrahi (submitted) found larger priming effects (40ms for agents, 44ms for patients) using a standard LDT task, which frequently observes larger effects than the less demanding pronunciation tasks (Neely, 1991). Using semantic decision tasks, which require conscious semantic processing, Hare et al. (2009) also observed larger priming effects (3237ms)
53 of event and location nouns on agents and patients. Although the variability in these results make strong claims regarding the relative contributions of automatic and attentional factors problematic, the smaller but significant priming effect observed in this study suggests that the continuous LDT paradigm more effectively isolated automatic processing, particularly when the longer average SOA in this study (which usually results in larger priming effects in paired prime target designs) is taken into account. R esearch Q uestion 3 RQ3 addressed the difference between the priming effect of two related primes (the RR condition) compared to one related prime followed by an unrelated noun (the RU condition). The effect of one related noun prime is significantly less than two related primes, when there is a second unrelated prime intervening. These results show an increased priming effect of the RR condition over RU that is very similar to the priming of RR vs. UU (less than 1ms different for either participant or item results). Because the RR vs. RU comparison in RQ3 was significant while the RR vs. UR comparison in RQ2 was not, it is possible that the time taken for presentation of prime 2 was sufficiently long for any preactivation from prime 1 to decay. The ordering of agent and patient primes was balanced across triads, so that thematic r ole order can be eliminated as a factor in the different results seen in RQ2 and RQ3. These results suggest that a single related noun prime is sufficient to account for the observed facilitation in healthy young adults, while preactivation decays too rapidly to facilitate a related verb in the presence of an intervening unrelated noun. Khader et al. (2003) also failed to observe a priming effect of a single related prime in the first position of a multiple prime task. However, Balota & Paul (1996) did observe priming effects of single related nouns on noun targets in all of their relatedness and task conditions, suggesting that the influence of the first prime in a multiple priming paradigm cannot be
54 disregarded. In Balota and Pauls (1996) converging prime paradigm, each prime activated different conceptual aspects of its prime (e.g. lion and stripes activate distinct category and visual conceptual features of the target tiger ). In contrast, the related primes in Khader et al. (2003) and in the pr esent study are related via common situation schemas. It is possible that an intervening, unrelated thematic role filler interfered with the activation of situation schemas, while presentation of intervening unrelated nouns in Balota and Paul (1996) did not interfere with activation of conceptual features of concrete nouns. Ferretti et al.s (2001) previous finding that agents and patients did not prime related verbs when used in an incongruous context suggests the possibility that exposure to semantic or syntactic information that contradicts realworld knowledge about common situation schemas may have an interference effect on activation of concepts related in this way, while the absence of syntactic context in the present study supports a semantic basis for such an effect. General Discussion Balota & Paul (1996) and Khader et al. (2003) both observed additive priming effects of multiple primes, when the pairings of primes and targets were explicit. Comparing the results of these two studies, longer SOAs and increased semantic processing demands (using relatedness judgment tasks that demand conscious, attentional processing of both primes and targets) were both potential contributors to a large boost in additive priming, and both have previously been ass ociated with greater attentional processing (Neely, 1991). The present study had fairly long average SOAs (1,432ms) similar to Khader et al. (2003) (1300ms). However, no additive facilitation was observed in this study, suggesting that the continuous lis t paradigm reduces the potential of strategic processing sufficiently to observe automatic processing despite the necessity of longer SOAs. The absence of additive facilitation in the present study strongly supports the hypothesis that taskspecific atten tional factors contributed to previously observed
55 additive priming effects using different multiple priming designs (Balota & Paul, 1996; Khader et al. 2003). Nonetheless, the smaller, statistically significant semantic priming effect observed in the cur rent study supports the argument that automatic factors do influence the activation of verbs in the context of exposure to multiple, thematically related nouns. Interestingly, there is an apparent (though not universal) trend toward faster RTs in continuous LDT compared to standard LDT (McNamara & Altarriba, 1988; Perea & Rosa, 2002). The available data do not support any conclusions, but there are several possible explanations for slower decisions in standard paired presentations compared to continuous list designs, where a response is required to primes. Silent reading of primes followed by lexical decisions on targets could involve an unintended element of taskswitching, which would increase the cognitive demands of the task. Alternatively, by requi ring a behavioral response to the prime, participants in a continuous list task signal that processing has occurred. In a standard task, although there is no behavioral measure of silent reading, processing time is still required, and ongoing processing of primes may delay processing of targets, adding to reaction times. In relation to the interference effects seen in previous pictureword interference and blocked semantic naming findings, in which exposure to closely related concepts (typically within c ategory) slows responses to following targets, it is interesting that no interference effect of two related noun primes was observed in any of the condition mean comparisons in this experiment (Sailor et al., 2008; Schnur et al., 2006). While no additive facilitation was seen for two thematically related primes compared to single primes, these results suggest that, at least in neurologically healthy young adults, there is no interference in verb processing from immediately preceding, multiple activations o f nouns all fitting within the kind of real world situation schema that forms the basis of natural discourse.
56 The present study supports the conclusion that automatic processing contributed to previously observed semantic priming effects of multiple primes but that taskspecific attentional factors that were greatly reduced or absent in this experiment also contributed, resulting in additive priming effects of multiple primes not observed in this study. When primes were explicitly paired with targets and conscious attention to the relationship between them was demanded, the addition of a second related prime resulted in an increase in the priming effect compared to single related primes (Balota & Paul, 1996; Khader et al ., 2003). The additive effects of a second related prime observed in Balota and Paul (1996) and in Khader et al. (2003) were likely boosted when the nature of the relationship between primes and targets was clear enough to allow for taskspecific strategic processing. With the prime prime target relationships of experimental interest sufficiently disguised by the continuous list design in this study, there is a greatly reduced likelihood that expectancy generation extending over both primes to the target was a significant factor in the observed priming effect. The priming effect of thematic role nouns on related verbs observed in this study is believed to be the result of automatic semantic processing suggesting a direct relationship between conceptual representations of agents and patient s to verbs that is not dependent on syntactic context or conscious memory of common situations. Future Directions and Clinical Implications More research is needed to better understand the implications of the current findings with respect to our broader un derstanding of the semantic relationships of verbs and their thematic roles. The selection of target verbs in this study was limited to 2 place thematic structures, which require two thematic role fillers to fill their simplest permitted syntactic frames (NP V NP). These findings might not extend to fillers of thematic roles that are optional in these verbs, such as instruments and locations. It is possible that the absence of additive priming in the
57 present study occurred because the agent and patient r oles activated similar conceptual features of the target verbs, and that other thematic roles could activate more and different features in the conceptual representation of verbs that may produce additive priming. A neutral condition using generic nouns ( i.e. someone or something) in addition to the related and unrelated prime conditions included in this study would test the proposed explanation for the failure to observe a priming effect of single related primes in the first position that the lack of a pr iming effect in the RU condition was related to interference from intervening incongruous role fillers. Nor can these findings be applied to all verbs, since the thematic role properties used to restrict selection of verbs in this study are not shared by all verb classes. Given previous sentence processing studies which have found differences in reaction times to 2place and 3 place verbs, i t is difficult to predict the possible effects of multiple primes on verbs with more complex stored lexical represen tations in single word semantic priming, where there is an absence of syntactic context ( Ahrens 2003; Shapiro et al., 1991) D ifferent priming pattern s might be seen in other populations as well. It is conceivable, though by no means certain, that mult iple primes could provide a boost in activation in older and impaired adults, leading to more robust semantic priming effects. Semantic priming effects have been demonstrated in healthy older adults, with typically slower reaction times, but sometimes lar ger priming effects (Bennet & McEvoy, 1999). Adults with aphasia have also demonstrated priming effects, though results are variable (Schnur, Schwartz, Brecher & Hodgson, 2006; Srs, Cornelissen, Laine & Salmelin, 2003). This is not surprising given the wide variability of lesion sites and language deficits seen in the population of peopl e with aphasia. If additive priming can be observed in older adults or people with aphasia, a multiple priming paradigm could potentially be used to guide stimuli and task selections in the design of future treatment studies
58 Increasing our understanding of how multiple sequential primes influence lexical retrieval could also someday be used to expand on existing diagnostic measures and interpretations in aphasia Inc reasing our knowledge of the semantic relationships between nouns and verbs and their neural bases is necessary to improving our understanding of the effects of aging and injury, and is critical to improving the assessment and treatment of sentence production deficits in aphasia (e.g., Edmonds, Nadeau, & Kiran, 2009). Treatment of aphasia has historically focused on naming (especially within categories) (see Nickels, 2002 for a review) or on production of syntactic structures (e.g., Thompson, Shapiro, Kira n, & Sobecks, 2003), with little interaction between them. Neither skill alone supports the complex task of conveying a meaningful message in a conversational context. Integrating thematic role and situation schema knowledge into the treatment of aphasia has the potential to better meet functional communication needs that significantly impact the lives of the individuals and families affected by this disorder.
59 APPENDIX A LIST OF EXPERIMENTAL STIMULI Related Prime 1 Related Prime 2 Target Unrelated Prime 1 Unrelated Prime 2 Role Order defendant prosecutor accusing spectator vegetarian PA traitor country betraying builder evening AP couple anniversary celebrating master examination AP dog bone chewing boy coal AP wood lumberjack chopping dirt cartooni st PA evidence detective discovering division architect PA ambulance paramedic driving appliance biologist PA tenant landlord evicting broker guardian PA electrician lights fixing participant sphere AP shuttle astronaut flying luggage machinist PA doctor patient healing father teacher AP host guest inviting ward clerk AP grandma sweater knitting rancher cottage AP math student learning mall captain PA boss employee managing aunt composer AP baker dough mixing widow sheet AP toilet plumber plungi ng closet athlete PA priest sermon preaching knight locker AP car tourist renting box dentist PA cat fireman rescuing cow nominee PA dice gambler rolling lime florist PA maid floor scrubbing fool scale AP waitress customer serving princess engineer A P soldier enemy shooting pianist member AP artist portrait sketching bishop hardware AP pig butcher slaughtering rat actress PA thief money stealing clown water AP ball pitcher throwing clay scholar PA movie audience watching arrow chairman PA novel author writing motor cowboy PA Primes 1 and 2 varied between related and unrelated by experimental condition. The order of presentation of agents and patients in experimental triads was balanced across patient agent (PA) and agent patient (AP). The rol e order of primes was the same across all conditions for each target verb.
60 APPENDIX B RELATEDNESS RATINGS FOR EXPERIMENTAL TRI ADS Agent Verb Patient Condition Mean prosecutor accusing defendant AP 6.87 prosecutor accusing spectator AU 3.73 vegeta rian accusing defendant UP 1.73 vegetarian accusing spectator UU 2.40 traitor betraying country AP 6.47 traitor betraying evening AU 1.60 builder betraying country UP 1.60 builder betraying evening UU 1.33 couple celebrating anniversary A P 6.80 couple celebrating examination AU 2.64 master celebrating anniversary UP 2.87 master celebrating examination UU 2.33 dog chewing bone AP 6.53 dog chewing coal AU 1.93 boy chewing bone UP 2.00 boy chewing coal UU 1.33 lumberja ck chopping wood AP 6.60 lumberjack chopping dirt AU 2.47 cartoonist chopping wood UP 1.93 cartoonist chopping dirt UU 1.40 detective discovering evidence AP 6.71 detective discovering division AU 2.53 architect discovering evidence UP 2. 20 architect discovering division UU 1.93 paramedic driving ambulance AP 6.93 paramedic driving appliance AU 1.67 biologist driving ambulance UP 1.47 biologist driving appliance UU 1.33 landlord evicting tenant AP 6.29 landlord evictin g broker AU 2.93 guardian evicting tenant UP 2.53 guardian evicting broker UU 2.07 electrician fixing lights AP 6.67 electrician fixing sphere AU 2.27 participant fixing lights UP 2.73 participant fixing sphere UU 1.53 astronaut flying shuttle AP 6.87 astronaut flying luggage AU 2.20 machinist flying shuttle UP 2.86 machinist flying luggage UU 1.53 doctor healing patient AP 6.73 doctor healing teacher AU 4.40 father healing patient UP 3.47 father healing teacher UU 2.07 host inviting guest AP 6.67 host inviting clerk AU 3.00 ward inviting guest UP 2.53 ward inviting clerk UU 2.27
61 Agent Verb Patient Condition Mean grandma knitting sweater AP 6.27 grandma knitting cottage AU 2.07 rancher knitt ing sweater UP 1.80 rancher knitting cottage UU 1.20 student learning math AP 6.53 student learning mall AU 1.73 captain learning math UP 2.33 captain learning mall UU 1.27 boss managing employee AP 6.87 boss managing composer AU 3.27 aunt managing employee UP 2.93 aunt managing composer UU 1.80 baker mixing dough AP 6.73 baker mixing sheet AU 2.07 widow mixing dough UP 2.93 widow mixing sheet UU 1.47 plumber plunging toilet AP 6.93 plumber plunging closet AU 1.5 7 athlete plunging toilet UP 1.87 athlete plunging closet UU 1.27 priest preaching sermon AP 6.87 priest preaching locker AU 1.40 knight preaching sermon UP 2.07 knight preaching locker UU 1.14 tourist renting car AP 6.53 tourist re nting box AU 1.73 dentist renting car UP 3.60 dentist renting box UU 2.00 fireman rescuing cat AP 6.40 fireman rescuing cow AU 2.93 nominee rescuing cat UP 2.07 nominee rescuing cow UU 1.47 gambler rolling dice AP 6.64 gambler rolli ng lime AU 1.53 florist rolling dice UP 1.47 florist rolling lime UU 1.73 maid scrubbing floor AP 6.47 maid scrubbing scale AU 3.07 fool scrubbing floor UP 3.67 fool scrubbing scale UU 1.73 waitress serving customer AP 6.87 waitress serving engineer AU 4.93 princess serving customer UP 1.40 princess serving engineer UU 1.40 soldier shooting enemy AP 6.87 soldier shooting member AU 3.67 pianist shooting enemy UP 1.40 pianist shooting member UU 1.47
62 Agent Verb Pa tient Condition Mean artist sketching portrait AP 6.40 artist sketching hardware AU 2.60 bishop sketching portrait UP 1.67 bishop sketching hardware UU 1.20 butcher slaughtering pig AP 6.40 butcher slaughtering rat AU 3.20 actress slau ghtering pig UP 1.67 actress slaughtering rat UU 1.71 thief stealing money AP 6.73 thief stealing water AU 3.36 clown stealing money UP 3.07 clown stealing water UU 2.47 pitcher throwing ball AP 6.87 pitcher throwing clay AU 2.13 sc holar throwing ball UP 3.33 scholar throwing clay UU 1.93 audience watching movie AP 6.53 audience watching arrow AU 2.60 chairman watching movie UP 3.40 chairman watching arrow UU 1.27 author writing novel AP 6.86 author writing motor AU 1.67 cowboy writing novel UP 1.93 cowboy writing motor UU 1.20 Average AP: 6.66 SD 0.20 Average AU: 2.56 SD 0.88 Average UP: 2.35 SD 0.72 Average mixed conditions (AU and UP): 2.46 SD 0.80 Average UU: 1.64 SD 0.40 Mean r atings from the relatedness judgment survey. The survey was completed by 15 participants who rated the commonality of the scenario created by each triad, by circling a number on a 7point ordinal scale where 1=not common at all, and 7=highly common. AP=r elated agent related patient condition; AU=related agent unrelated patient condition; UP=unrelated agent, related patient condition; UU=unrelated agent unrelated patient condition. SD=standard deviations.
63 APPENDIX C PSYCHOLINGUISTIC DAT A ON STIMULI ITE MS Verb V AOA V KFFRQ V NLET V NPHN V NSYL V CNC V FAM V IMG Mean 306.8 20.52 8.03 6.43 2.33 451 532.8 472.6 N= 5 29 30 30 30 3 5 5 accusing 8 8 6 3 betraying 1 9 8 3 celebrating 5 11 10 4 chewing 13 7 6 2 chopping 5 8 6 2 discovering 7 11 9 4 driving 53 7 6 2 evicting 8 7 3 fixing 11 6 6 2 flying 43 6 5 1 healing 6 7 5 2 inviting 8 8 7 3 knitting 286 1 8 5 2 583 501 578 learning 322 60 8 6 2 30 3 565 370 managing 8 8 8 3 mixing 10 6 6 2 plunging 406 3 8 8 2 preaching 17 9 7 2 454 441 renting 5 7 6 2 rescuing 2 8 7 3 rolling 19 7 5 2 514 434 scrubbing 3 9 7 2 serving 38 7 6 2 shooting 48 8 5 2 sketching 4 9 7 2 slaughtering 1 12 7 3 stealing 6 8 6 2 throwing 264 17 8 5 2 watching 76 8 6 2 writing 256 117 7 5 2 467 630 540 Normative ratings data are not frequently available fo r verbs in the present progressive ing form used in this study; ratings and written frequency data for the infinite forms of the target verbs was also collected. V=verb (in present progressive ing form). N=number of data points in mean. Psycholinguist ic data were collected from the MRC Psycholinguistic Database (Clark, 1997). AOA=Age of acquisition, range 100700, from Gilhooly & Logies (1980) normative data, mathematically transformed to match the 100 700 point range used for the other rated variable s. phonemes. NSYL=number of syllables. CNC=concreteness, FAM=familiarity, and IMG=imageability ratings on a range of 100700, derived by merging normative da ta from three sources, the Pavio Norms (an unpublished expansion of Pavio, Yuille & Madigan, 1968), Toglia and Battig (1978) and Gilhooly and Logie (1980) by a method described by Coltheart (1981).
64 Verb (Infinitive) Inf AOA Inf KFFRQ Inf CNC Inf FAM Inf IMG Mean 365 29.03 443.2 530.69 484.81 N= 6 29 15 16 16 accuse 10 betray 4 celebrate 4 chew 2 442 570 527 chop 317 3 555 487 575 discover 40 drive 105 evict fix 14 fly 33 525 537 58 2 heal 2 417 536 438 invite 11 knit 10 learn 84 370 552 361 manage 20 508 303 mix 13 plunge 406 5 396 441 548 preach 8 rent 425 21 437 547 469 rescue 367 15 373 532 456 roll 35 457 547 496 scrub 9 serve 107 shoot 27 445 536 494 sketch 411 16 535 452 510 slaughter 10 steal 5 363 562 448 throw 264 42 400 548 477 watch 81 487 576 525 write 106 446 560 548 Normative ratings data are not frequently available for verbs in the present progressive ing form used in this study; ratings and written frequency data for the infinite forms of the target verbs was also collected. Inf=infinite form of verb. N=number of data points in mean. Psycholinguistic data were collected f rom the MRC Psycholinguistic Database (Clark, 1997). AOA=Age of acquisition, range 100 700, from Gilhooly & Logies (1980) normative data, mathematically transformed to match the 100 700 point range used for the other rated variables. KFFREQ=written freque NSYL=number of syllables. CNC=concreteness, FAM=familiarity, and IMG=imageability ratings on a range of 100700, derived by merging normative data from three sources, t he Pavio Norms (an unpublished expansion of Pavio, Yuille & Madigan, 1968), Toglia and Battig (1978) and Gilhooly and Logie (1980) by a method described by Coltheart (1981).
65 Related Prime 1 P1 AOA P1 KFFRQ P1 NLET P1 NPHN P1 NSYL P1 CNC P1 FAM P1 IMG Mean 306.09 48.82 5.633 4.677 1.73 553.37 541.05 568 N= 11 28 30 30 30 19 20 19 defendant 6 9 9 3 traitor 2 7 5 2 467 467 447 couple 122 6 4 2 dog 169 75 3 3 1 610 598 636 wood 269 55 4 3 1 606 574 577 evidence 480 204 8 7 3 340 50 4 356 ambulance 6 9 9 3 595 499 627 tenant 5 6 6 2 electrician 11 10 4 shuttle 7 4 2 doctor 100 6 5 2 575 573 600 host 36 4 4 1 grandma 13 7 6 2 math 4 4 3 1 386 548 487 boss 20 4 3 1 552 574 554 baker 36 5 4 2 toilet 13 6 5 2 586 567 603 priest 411 16 6 5 1 561 484 568 car 197 274 3 2 1 622 634 638 cat 23 3 3 1 615 582 617 dice 14 4 3 1 487 maid 31 4 3 1 waitress 2 8 6 2 soldier 275 39 7 5 2 578 517 578 artist 57 6 5 2 554 547 600 pig 233 8 3 3 1 614 509 635 thief 322 8 5 3 1 519 529 529 ball 150 110 4 3 1 615 575 622 movie 386 29 5 4 2 590 523 571 novel 475 59 5 5 2 529 530 547 P1=related prime 1 N=number of data points in mean. Psycholingui stic data were collected from the MRC Psycholinguistic Database (Clark, 1997). AOA=Age of acquisition, range 100700, from Gilhooly & Logies (1980) normative data, mathematically transformed to match the 100 700 point range used for the other rated variables. phonemes. NSYL=number of syllables. CNC=concreteness, FAM=familiarity, and IMG=imageability ratings on a range of 100700, derived by merging normative data from three sources, the Pavio Norms (an unpublished expansion of Pavio, Yuille & Madigan, 1968), Toglia and Battig (1978) and Gilhooly and Logie (1980) by a method described by Coltheart (1981).
66 Related Prime 2 P2 AOA P2 KFFRQ P2 NLET P2 NPHN P2 N SYL P2 CNC P2 FAM P2 IMG Mean 374.4 58.67 7.18 5.833 2.3 529.88 531.06 541 N= 10 27 30 30 30 17 18 17 prosecutor 8 10 10 4 520 454 497 country 329 324 7 6 2 465 592 539 anniversary 21 11 9 5 bone 33 4 3 1 588 541 567 lumberjack 10 9 3 detective 377 52 9 8 3 505 509 524 paramedic 9 9 4 landlord 12 8 6 2 lights 47 6 4 1 astronaut 2 9 8 3 patient 366 86 7 6 2 487 538 526 guest 39 5 4 1 519 560 497 sweater 14 7 5 2 student 481 131 7 8 2 549 632 603 employee 24 8 6 3 dough 369 13 5 2 1 627 474 558 plumber 4 7 5 2 sermon 12 6 5 2 tourist 16 7 6 2 533 536 577 fireman 1 7 6 3 504 gambler 7 6 2 floor 158 5 3 1 559 551 544 customer 369 27 8 7 3 505 549 488 enemy 367 88 5 5 3 434 523 497 portrait 17 8 6 2 570 427 565 butcher 8 7 5 2 556 473 596 money 247 265 5 4 2 574 631 604 pitcher 21 7 5 2 audience 425 115 8 6 2 515 511 555 author 414 46 6 3 2 502 554 460 P2=related pr ime 2 N=number of data points in mean. Psycholinguistic data were collected from the MRC Psycholinguistic Database (Clark, 1997). AOA=Age of acquisition, range 100700, from Gilhooly & Logies (1980) normative data, mathematically transformed to match t he 100700 point range used for the other rated variables. phonemes. NSYL=number of syllables. CNC=concreteness, FAM=familiarity, and IMG=imageability rating s on a range of 100700, derived by merging normative data from three sources, the Pavio Norms (an unpublished expansion of Pavio, Yuille & Madigan, 1968), Toglia and Battig (1978) and Gilhooly and Logie (1980) by a method described by Coltheart (1981).
67 Unrelated Prime 1 U1 AOA U1 KFFRQ U1 NLET U1 NPHN U1 NSYL U1 CNC U1 FAM U1 IMG Mean 390.22 40 5.633 4.577 1.73 550.92 511.56 542.64 N= 9 30 30 30 30 25 25 25 spectator 9 9 8 3 builder 29 7 5 2 532 554 551 master 72 6 5 2 498 495 495 boy 242 3 2 1 609 606 618 dirt 43 4 3 1 564 571 547 division 417 107 8 7 3 332 518 404 appliance 5 9 7 3 558 493 554 broker 1 6 5 2 participant 575 4 11 10 4 439 430 370 luggage 10 7 5 2 father 183 6 4 2 594 591 646 ward 25 4 3 1 rancher 15 7 6 2 mall 589 3 4 3 1 459 330 383 aunt 233 22 4 3 1 564 554 567 widow 426 26 5 4 2 547 485 505 closet 16 6 6 2 599 540 525 knight 18 6 3 1 579 505 608 box 192 70 3 4 1 597 599 591 cow 29 3 2 1 621 529 632 lime 378 13 4 3 1 590 447 563 fool 37 4 3 1 354 551 436 princess 10 8 7 2 568 502 547 pianist 14 7 7 2 565 473 608 bishop 18 6 5 2 587 467 524 rat 6 3 3 1 624 548 588 clown 3 5 4 1 627 511 589 clay 358 100 4 3 1 606 455 575 arrow 14 5 3 2 595 490 619 motor 344 56 5 4 2 565 545 521 U1=unrelated prime 1 N=number of data points in mean. Psycholinguistic data were collected from the MRC Psycholinguistic Database (Clark, 1997). AOA=Age of acquisition, range 100700, from Gilhooly & Logies (1980) normative data, mathematically transformed to match the 100 700 point range used for the other rated variables. phonemes. NSYL=number of syllables. CN C=concreteness, FAM=familiarity, and IMG=imageability ratings on a range of 100700, derived by merging normative data from three sources, the Pavio Norms (an unpublished expansion of Pavio, Yuille & Madigan, 1968), Toglia and Battig (1978) and Gilhooly an d Logie (1980) by a method described by Coltheart (1981).
68 Unrelated Prime 2 U2 AOA U2 KFFRQ U2 NLET U2 NPHN U2 NSYL U2 CNC U2 FAM U2 IMG Mean 347 46.07 7.2 5.8 2.27 532.89 537.58 545.28 N= 11 30 30 30 30 18 19 18 vegetarian 1 10 10 4 evenin g 303 133 7 5 2 439 630 559 examination 458 29 11 11 5 464 570 575 coal 278 32 4 3 1 584 513 581 cartoonist 3 10 8 3 architect 22 9 7 3 biologist 2 9 9 4 guardian 9 8 6 2 sphere 22 6 3 1 489 457 562 machinist 5 9 8 3 teacher 247 80 7 5 2 569 599 575 clerk 34 5 4 1 cottage 19 7 5 2 593 543 607 captain 85 7 6 2 534 498 497 composer 461 31 8 7 3 487 448 467 sheet 251 45 5 3 1 608 616 594 athlete 428 9 7 5 2 545 482 591 locker 9 6 4 2 586 5 38 569 dentist 12 7 7 2 607 563 622 nominee 3 7 6 3 florist 1 7 7 2 scale 397 60 5 4 1 475 523 463 engineer 449 42 8 6 3 531 514 495 member 392 137 6 5 2 455 573 399 hardware 11 8 5 2 560 496 576 actress 6 7 6 2 water 1 53 442 5 4 2 616 641 632 scholar 15 7 5 2 450 489 451 chairman 67 8 6 2 cowboy 16 6 4 2 521 U2=unrelated prime 2 N=number of data points in mean. Psycholinguistic data were collected from the MRC Psycholinguistic Database (Clark, 199 7). AOA=Age of acquisition, range 100700, from Gilhooly & Logies (1980) normative data, mathematically transformed to match the 100 700 point range used for the other rated variables. of letters. NPHN=number of phonemes. NSYL=number of syllables. CNC=concreteness, FAM=familiarity, and IMG=imageability ratings on a range of 100700, derived by merging normative data from three sources, the Pavio Norms (an unpublished expansion of Pavio Yuille & Madigan, 1968), Toglia and Battig (1978) and Gilhooly and Logie (1980) by a method described by Coltheart (1981).
69 APPENDIX D COMPLETE TEXT OF EXP ERIMENT INSTRUCTIONS The following text includes all instructions given to participants in this p riming experiment during the instruction and practice sections of the session. Instructions, practice, and the priming experiment were all presented on a laptop computer using DirectRT software, as described in the methods section, chapter 3. Welcome! In this experiment, you will read strings of letters and decide whether or not they are real English words. Press enter to continue. Find the keys marked Y (yes) and N (no) and place the index and middle fingers of your right hand over them. You'll use these keys to indicate your decisions. Be accurate in your answers and type your responses as quickly as you can after you decide. Ready? Let's try a few examples. house o Correct! HOUSE is a real English word. Let's try another. Hit enter to continue. or o O ops! HOUSE is a real English word. You typed N for no. The correct response is Y for yes. Let's try that one again. Hit enter to continue. smoufe o Correct! SMOUFE is NOT a real English word. Let's try another. Hit enter to continue. or o Oops! SMOUFE is NOT a real English word. You typed Y for yes. The correct response is N for no. Let's try that one again. Hit enter to continue. walking o Correct! WALKING is a real English word. Let's try another. Hit enter to continue. or o Oops! WALKING is a real Engl ish word. You typed N for no. The correct response is Y for yes. Let's try that one again. Hit enter to continue. kalping o Correct! KALPING is NOT a real English word. You've got it! Let the examiner know if you have any questions, and press enter when y ou're ready to continue. or o Oops! KALPING is NOT a real English word. You typed Y for yes. The correct response is N for no. Let's try that one again. Hit enter to continue.
70 You'll now complete some practice items. Continue to use the Y and N keys to indicate your response. The items will now more forward automatically if you're correct, but you will get feedback if you make a mistake or respond too fast or too slow. Press enter when you're ready to begin the practice items. End of practice items. P ress enter to continue. You're now ready to begin the experiment. You will no longer receive feedback about your performance, but continue to the next item automatically after each response. Continue to respond accurately and to type your responses as quickly as you can. If you have any problems, just stop typing responses and ask the examiner. There are three parts to this experiment, and you'll take two five minute breaks between them. After you finish you'll have an opportunity to talk with the exam iner about the experiment. Any questions before you begin? Press enter to begin. Take a five minute break. Press enter to begin the next part of the experiment. Take a five minute break. Press enter to begin the last part of the experiment. That's all! THANKS for your participation!
71 LIST OF REFERENCES Ahrens, K. (2003). Verbal integration: The interaction of participant roles and sentential argument structure. Journal of Psycholinguistic Research, 32, 497516. Ahrens, K., & Swinney, D. (1995). Participant roles and the processing of verbs during sentence comprehension. Journal of Psycholinguistic Research, 24, 533547. Balota, D. A., Black, S. R., & Cheney, M. (1992). Automatic and attentional priming in young and older adults: Reevaluation of the twoprocess model. Journal of Experimental Psychology: Human Perception and Performance, 18, 485502. Balota, D. A., & Paul, S. T. (1996). Summation of activation: Evidence from multiple primes that converge and diverge within semantic memory. Journal of Experi mental Psychology: Learning, Memory, and Cognition, 22, 827845. Bennett, D. J., & McEvoy, C. L. (1999). Mediated priming in younger and older adults. Experimental Aging Research, 25, 141159. Clark, C. (1997). MRC psycholinguistic database (Version 2.00) [Machine usable dictionary]. Available from University of Western Australia Web site, http://www.psy.uwa.edu.au/mrcdatabase/uwa_mrc.htm Coltheart, M. (1981). MRC Psycholinguistic Database User Manual: Version 1. Retrieved February 9, 2011 from http://www .psych.rl.ac.uk/User_Manual_v1_0.html. DirectRT Research Software (2004). New York, NY: Empirisoft. Edmonds, L. A., & Mizrahi, S. (submitted). Online priming of agent and patient thematic roles and related verbs in younger and older adults. Edmonds, L. A ., Nadeau, S. E., & Kiran, S. (2009). Effect of verb network strengthening treatment (VNeST) on lexical retrieval of content words in sentences in persons with aphasia. Aphasiology, 23, 402424. Federmeier, K. D., Segal, J. B., Lombrozo, T., & Kutas, M. (2000). Brain responses to nouns, verbs and class ambiguous words in context. Brain, 123, 25522566. Ferretti, T. R., McRae, K., & Hatherell, A. (2001). Integrating verbs, situation schemas, and thematic role concepts. Journal of Memory and Language, 44, 516547. Field, A. (2009). Discovering statistics using SPSS: Introducing statistical methods. (3rd ed.) London: SAGE Publications., Ltd. Frequency analysis of English usage: lexicon and grammar. Boston: Houghton Mifflin.
72 Gilhooly, K.J. and Logie, R.H. (1980). Age of acquisition, imagery, concreteness, familiarity and ambiguity measures for 1944 words. Be haviour Research Methods and Instrumentation, 12, 395427. Grondin, R., Lupker, S. J., & McRae, K. (2009). Shared features dominate semantic richness effects for concrete concepts. Journal of Memory and Language, 60, 119. Hare, M., Jones, M., Thomson, C., Kelly, S., & McRae, K. (2009). Activating event knowledge. Cognition, 111, 151167. Holm, S. (1979). A simple sequentially rejective multiple test procedure. Scandinavian Journal of Statistics, 6, 6570. Jones, L. L. (2010). Pure mediated priming: A retr ospective semantic matching model. Journal of Experimental Psychology: Learning, Memory and Cognition, 36, 135146. Khader, P., Scherag, A., Streb, J., & Rsler, F. (2003). Differences between noun and verb processing in a minimal phrase context: a semanti c priming study using event related brain potentials. Cognitive Brain Research, 17, 293313. Kipper, K., Korhonen, A., Ryant, N., & Palmer, M. (2008). A large scale classification of English verbs. Language Resources and Evaluation Journal, 42, 2140. Kou nios, J., Green, D. L., Payne, L, Fleck, J. I., Grondin, R., & McRae, K. (2009). Semantic richness and the activation of concepts in semantic memory: Evidence from event related potentials. Brain Research, 1282, 95102. Computational analysis of present day American English. Providence: Brown University Press. Levin, B. (1993). English verb classes and alternations: A preliminary investigation. Chicago: University of Chicago Press. McR ae, K., & Boisvert, S. (1998). Automatic semantic similarity priming. Journal of Experimental Psychology: Learning, Memory, and Cognition, 24, 558572. McRae, K., Ferretti, T. R., & Amyote, L. (1997). Thematic roles as verbspecific concepts. Language and Cognitive Processes, 12, 137176. McRae, K., Hare, M., Elman, J. L., & Ferretti, T. (2005). A basis for generating expectancies for verbs from nouns. Memory & Cognition, 33, 11741184. McNamara, T. P., & Altarriba, J. (1988). Depth of spreading activation revisited: Semantic mediated priming occurs in lexical decisions. Journal of Memory and Language, 27, 545559.
73 Merritt, P., DeLosh, E. L., & McDaniel, M. A. (2006). Effects of word frequency on individual item and serial order retention: Tests of the orde r encoding view. Memory & Cognition, 34, 16151627. Microsoft. (2007). Microsoft Office Access (Version 2007) [Computer software]. Redmond, WA: Microsoft. Moss, H. E., Ostrin, R. K., Tyler, L. K., & MarslenWilson, W. D. (1995). Accessing different types of lexical semantic information: Evidence from priming. Journal of Experimental Psychology: Learning, Memory, and Cognition, 21, 863883. Neely, J. H. (1991). Semantic priming effects in visual word recognition: A selective review of current findings and theories. In D. Besner & G. W. Humphreys (Ed.), Basic processes in reading (pp. 264336). Hillsdale, NJ: Lawrence Erlbaum Associates. Nelson, D. L., McEvoy, C. L., & Schreiber, T. A. (1998). The University of South Florida word association, rhyme, and word fragment norms. Retrieved January 21, 2011 from http://web .usf.edu/FreeAssociation/. Nickels, L. (2002). Therapy for naming disorders: Revisiting, revising, and reviewing. Aphasiology, 16, 935979. Palmer, M., & Kipper, K. (2010). Unified verb index [Data file] Available from University of Colorado Web site, http://verbs.colorado.edu/verb index/vn/class h.php.Paivio A., Yuille, J.C. and Madigan, S.A. (1968). Concreteness, imagery and meaningfulness values for 925 words. Journal of Experimental Psychology Monograph Supplement, 76 (3, part 2). Perea, M., & Rosa, E. (2002). The effects of associative and semantic priming in the lexical decision task. Psychological Research, 66, 180 194. Raczaszek Leonardi, J., Shapiro, L. P., Tuller, B., & Kelso, J. A. S. ( 2008). Activating basic category exemplars in sentence contexts: A dynamic account. Journal of Psycholinguistic Research, 37, 87113. Rastle, K., Harrington, J., & Coltheart, M. (2002). 358,534 nonwords: The ARC Nonword Database. Quarterly Journal of Expe rimental Psychology, 55, 13391362. Sailor, K., Brooks, P. J., Bruening, P. R., Seiger Gardner, L., & Guterman, M. (2008). Exploring the time course of semantic interference and associative priming in the pictureword interference task. The Quarterly Jour nal of Experimental Psychology, 62, 789801. Schnur, T. T., Schwartz, M. F., Brecher, A., & Hodgson, C. (2006). Semantic interference during blockedcyclic naming: Evidence from aphasia. Journal of Memory and Language, 54, 199227.
74 Setola, P., & Reilly, R. G. (2005). Words in the brains language: An experimental investigation. Brain and Language, 94, 251259. Shapiro, L. P., Brookins, B., Gordon, B., & Nagel, N. (1991). Verb effects during sentence processing. Journal of Experimental Psychology: Learning, Memory, and Cognition, 17, 983996. Srs, P., Cornelissen, K., Laine, M., & Salmelin, R. (2003). Naming actions and objects: Cortical dynamics in healthy adults and in an anomic patient with a dissociation in action/object naming. NeuroImage, 19, 17871801. Thompson, C. K., Shapiro, L. P., Kiran, S., & Sobecks, J. (2003). The role of syntactic complexity in treatment of sentence deficits in agrammatic aphasia: The complexity account of treatment efficacy (CATE). Journal of Speech, Language, and Hearing Re search, 46, 591607. Toglia, M.P., & Battig, W.R. (1978). Handbook of Semantic Word Norms. New York: Erlbaum. Vigliocco, G., Vinson, D. P., Lewis, W., & Garrett, M. F. (2004). Representing the meanings of object and action words: The featural and unitary semantic space hypothesis. Cognitive Psychology, 48, 422488. Yap, M. J., Tse, Chi Shing, & Balota, D. A. (2009). Individual differences in the joint effects of semantic priming and word frequency revealed by RT distributional analyses: The role of lexical integrity. Journal of Memory and Language, 61, 303325.
75 BIOGRAPHICAL SKETCH Stacey Herlofsky was born in Minneapolis Minnesota She graduated from Minneapolis Edison Senior High School in 1995. She received a Bachelor of Arts degree in linguistics a nd anthropology from the University of Chicago in 1999. Stacey began coursework for a degree in communication sciences and disorders at the University of Florida in fall 2008. After completing her Master of Arts degree in May 2011 she began a clinical fellowship in the field of speech language pathology