Brain and Language 77, 45–59 (2001) doi:10.1006/brln.2000.2422, available online at http://www.idealibrary.com on Acquisition of New ‘‘Words’’ in Normal Subjects: A Suggestion for the Treatment of Anomia Anna Basso,*,† Paola Marangolo,† Fabrizio Piras,† and Claudia Galluzzi† *Clinica Neurologica, Università di Milano, Milan, Rome; and †IRCCS S. Lucia, Milan, Rome Published online February 15, 2001 The study explores the efficacy of three learning methods in normal controls. Thirty subjects, randomly assigned to the repetition, reading aloud, or orthographic cueing method, were asked to learn 30 new ‘‘words’’ (legal nonwords arbitrarily assigned to 30 different pictures); 30 further new ‘‘words’’ were used as controls. Number of trials to criterion was significantly lower, and number of words remembered at follow-up was significantly higher for the orthographic cueing method. Two aphasic patients with damage to the output lexicons were also rehabilitated with the same three methods. In both patients the orthographic cueing method was significantly more efficacious. The differences in learning efficacy of the three methods are discussed. 2001 Academic Press INTRODUCTION When we speak, word retrieval is usually a quick and easy process albeit wordfinding difficulties may also occur in normal speakers. At times, proper names or precise words may in fact be difficult to retrieve even if we know them. In aphasic patients difficulty in word retrieval is the most pervasive symptom of language breakdown, and naming disorders may result in a wide variety of errors due to damage to different stages in the process of naming. The similarities between aphasic errors and normal slips of the tongue have been noted by many authors. Freud claimed that ‘‘the paraphasia in aphasic patients does not differ from the incorrect use and the distortion of words which the healthy person can observe in himself in states of fatigue or divided attention’’ (1953, p. 13). The pervasiveness of word-finding difficulties has motivated several studies devoted to the management of the deficit and its effectiveness. Group studies suggest that word-finding deficits in general can be ameliorated (Hillis, 1989; Howard, Patterson, Franklin, Orchard-Lisle, & Morton, 1985a, 1985b; Marshall, Pound, WhiteThompson, & Pring, 1990; Myers-Pease & Goodglass, 1978) but in group studies it is difficult to evince what treatment has been useful for what type of patients. Error types in normal subjects and aphasic patients have provided important clues to the architecture of the normal lexical processing system and many authors have recently proposed relying on cognitive neuropsychological models to reach a functional diagnosis, which should then be used as a guide to aphasia therapy (Behrmann, 1987; Byng, 1988; Hillis & Caramazza, 1987). Theoretically based treatments have been published presenting cases of patients whose deficits were identified relative to a Address correspondence to Anna Basso, Neurological Clinic Milan University, Via F. Sforza 35, 20122 Milan, Italy. Fax: 39 02 546 3834. E-mail: [email protected] 45 0093-934X/01 $35.00 Copyright 2001 by Academic Press All rights of reproduction in any form reserved. 46 BASSO ET AL. FIG. 1. Functional architecture of the lexical-semantic system and the sublexical-conversion mechanisms. functional model of single word processing and rehabilitated following the cognitive neuropsychological approach (Howard et al., 1985b; Miceli, Amitrano, Capasso & Caramazza, 1996). The background model we will refer to has the functional architecture shown in Fig. 1 (for a detailed description of the model see Caramazza & Hillis, 1990; Miceli, Giustolisi, & Caramazza, 1991). Briefly, the model assumes that written language does not depend on spoken language and makes a distinction between input and output word-form stores. The single semantic component is independently connected with the phonological and the orthographic input and output lexicons, which contain the phonological and the orthographic representations of words. The segmental processing of new words in reading aloud is based on grapheme-to-phoneme conversion rules, and in writing to dictation on phoneme-to-grapheme conversion rules; the seg- ACQUISITION OF NEW WORDS 47 mental processing in repetition is based on input to output phoneme conversion rules that are not represented in Fig. 1. Repetition, writing, and reading aloud of regular words can be performed by the dual activation of the lexical and the sublexical procedures. Finally, it is assumed that the lexical and sub-lexical procedures interact. This model predicts different error types in a naming task according to the site of the functional damage. In the present study, we are only interested in damage to the phonological output lexicon, and thus our focus will be on patients with a loss of information about the phonological representations and intact verbal semantics. In these patients, the complete semantic information fails to activate an unavailable (or inaccessible) lexical representation and patients produce either omissions, descriptions of the object to be named (circumlocutions), or partial phonological renditions of the target word. It has also been argued that semantic paraphasias can results from damage to the output lexicon (see, for instance, Caramazza & Hillis, 1990). The cognitive approach allows us to reach a precise functional diagnosis but it does not provide guidelines for the implementation of rehabilitation (for a discussion, see Caramazza, 1989). In fact the models fail to include many important aspects of the rehabilitation process, such as treatment techniques likely to modify the identified functional damage/s. To build a theory of rehabilitation, the hypothesis at the basis of the proposed intervention must clearly be made explicit and the proposed intervention should then be tested. It is our opinion that, short of counterevidence about its lack of effectiveness, therapy should be directly aimed at the functional damage. We also argue that, with some obvious limitations (such as those due to the general effect of the presence of brain damage, which obviously plays an important role), we can compare a normal subject performing a given task to a brain-damaged patient performing the same task. If this is the case, the techniques most successful in normal subjects for solving a given task can be used as a starting point for implementing a therapeutic method for patients with a functional damage which impairs the execution of the same task. In this study our aim is to identify the technique that is most successful in normal subjects’ learning of new words. When asked to learn words in an unknown second language or non words associated with a given picture, normal subjects have a complete semantic information about the concept. They cannot, however, activate the word in the output lexicon because, by definition, the phonological representation of the word is not there. Aphasic patients with an intact semantic component and damage to the phonological output lexicon would be in the same situation (except, as we pointed out earlier, for the consequences of brain damage that obviously play a role). If we can identify such a technique, the following rational step would be to see whether it is successful with patients with the same functional ‘‘damage’’ of normal subjects, namely damage to the phonological output lexicon with an intact semantic system. Tikofsky and coworkers (Tikofsky, 1971; Carson, Carson, & Tikofsky, 1968) have in fact reported data that indicate that relative to normal controls, aphasic patients as a group show reduced rates and amounts of verbal learning but that their performance patterns are quite similar. The paper reports on an experimental study in normal controls who where asked to learn new ‘‘words’’ associated with pictures. We decided to test three techniques normally used in aphasia therapy for the rehabilitation of anomia: reading aloud, repetition, and the use of a phonological cue. However, instead of the phonological cue we used an orthographic cue. We chose an orthographic instead of the more usual phonological cue for two main reasons. First, the ‘‘prosody’’ of the phonological cue is by itself an important cue; it can be produced with a rising tone that can provoke a quasi automatic response from the patient (see, for instance patient JCU who said 48 BASSO ET AL. tiger in front of the picture of a lion if cued with /t/; Howard & Orchard-Lisle, 1984) and we wanted to elicit an intentional response. Second, since the orthographic cue remains in view of the subject, he or she has longer to search for the target word. The orthographic cue can be used to implement oral naming because the orthographic output can ‘‘recirculate’’ in the lexical system. The written response can be converted by grapheme-to-phoneme conversion rules and produced orally. In order to study the efficacy of the three learning methods, we asked three groups of normal subjects to learn new words, each associated with a picture, and compared the efficacy of the three methods. We also tested two anomic patients, RF and MR, with damage to the output lexicons and intact conversion mechanisms necessary for reading aloud and repetition. EXPERIMENT Subjects Thirty healthy right-handed volunteers (18 men and 12 women) participated in the study. All subjects gave informed consent; they were native Italian speakers aged between 29 and 63 years (mean ⫽ 49.3, SD ⫽ 8.5) with 8 to 17 years of formal education (mean ⫽ 12.4, SD ⫽ 2.5). Subjects were randomly subdivided into three groups (10 subjects in each group) matched for age and educational level. Stimuli A set of 60 bisyllabic invented words was prepared. Words were construed in such a way as to have a very low level of similarity with Italian words, to include a consonant cluster, and to have a one-toone phoneme-to-grapheme correspondence. Thirty words were used for learning (experimental) and 30 as controls. Sixty pictures of low, medium, and high familiarity, belonging to different semantic categories (animals, tools, body parts, clothes, fruits and vegetables, and musical instruments) from the Snodgrass and Vanderwart’s (1980) picture-set, were selected. The 60 pictures were randomly matched to the invented words, one picture per word, with the same number of low, medium, and high familiarity pictures in the experimental and control group. The 60 new words and the associated Italian words are reported in the appendix. Procedure Subjects were tested individually. The experiment included three phases: a training phase, a learning phase, and a follow-up. The first two phases were run on two consecutive days, and the third a week later. Training. The procedure for the training phase was identical for the three learning methods and was as follows: the 60 pictures were presented one by one to the subject in pseudorandom order while the examiner clearly said the corresponding invented words. The subject was instructed to pay attention to each word-picture pair without repeating the word aloud. The pictures were presented three times and at the end of the third presentation, the subject was presented with a display of the 60 pictures and was required to point to the picture named by the examiner. If he or she pointed to an incorrect picture, the examiner pointed to the correct one saying the corresponding word and then went on to the following stimulus. The number of pictures correctly indicated by the subjects varied from 4 to 16 (mean ⫽ 9.8, SD ⫽ 3.3). On the second day, the subject was first presented once again with the display of the 60 pictures, and was required to point to the picture named by the examiner. If the subject pointed to an incorrect picture, the examiner did not comment on the error in any way and said the following word. The number of pictures correctly pointed to by the subjects varied from 3 to 20 (mean ⫽ 8.8, SD ⫽ 4.0). The 60 pictures were then presented one by one to the subject and he or she was asked to name them without any feedback; the number of correct responses was recorded (baseline). Learning. The learning phase followed immediately and was different for the three experimental groups. In the learning phase and in the subsequent naming condition the pictures were presented in pseudorandom order. For each group of subjects a different cueing method was used. The first group was asked to learn the words by repeating aloud the experimental stimulus presented by the examiner together with the corresponding picture. The examiner showed the pictures one by one and waited three 49 ACQUISITION OF NEW WORDS seconds for the subject to say the corresponding word. If he or she failed, the examiner said the word and the subject was required to repeat it. After presentation of the 30 experimental pictures, the subjects was presented with all the pictures (30 experimental and 30 control pictures) and asked to say the ‘‘name’’ of each picture. The number of correct experimental and control words was recorded. The whole naming procedure was then repeated until the subject correctly named all 30 experimental words (or for a maximum of 20 trials. All subjects, however, learned the 30 experimental words in less than 20 presentations and only one subject required as many as 10). Subjects in the second group were asked to learn the words by reading aloud the written word presented with the corresponding picture. The procedure was the same as for the repetition group. The third group was asked to learn the words using an orthographic cue. The examiner presented the pictures one by one and asked the subject to say the name; if he or she could not say it in three seconds, the examiner wrote the first letter of the corresponding word and waited 3 s for the subject to complete the word. If he or she could not, the same procedure was used for the following letters until the subject completed the written word and then said it. The procedure was then the same as for the reading and the repetition groups. Follow-up. A week later and without previous notice, the subjects were once again presented with the 60 pictures one by one and asked to name them without any feedback. STATISTICAL PROCEDURE AND RESULTS Table 1 reports for each of the 10 subjects of the repetition group the number of experimental and control pictures correctly named at baseline, the number of presentations necessary to learn the 30 experimental words, and the number of experimental and control pictures correctly named at follow-up. Tables 2 and 3 report the same data for the reading aloud and the orthographic cueing group, respectively. The graph in Fig. 2 illustrates the same data for the three groups. Number of presentations to criterion. For each subject in each experimental group the number of presentations necessary to learn the 30 experimental words was computed and the results of the three groups were compared. The data were analyzed using an Anova with one within–subject factor (cueing method: repetition, reading aloud, orthographic cue). The analysis revealed a significant difference between the three cueing methods: F ⫽ 19.528, df 2, 27; p ⫽ .0000. The mean number of presentations in learning the 30 experimental words with the orthographic cue was significantly lower (5.1) than the number of presentations in learning the experimental TABLE 1 Repetition Method: Number of Stimuli (Experimental and Control) Correctly Named by the Subjects at Baseline and Follow-Up, and Number of Presentations Necessary to Learn the 30 Experimental Words Baseline (stimuli) Subjects AA MB AP AM ED GB ADF CDF GL LL Mean (SD) Exp. Contr. 4 3 5 1 6 3 3 3 4 5 3.7 (1.4) 2 4 6 1 4 3 7 2 3 4 3.6 (1.8) Note. Exp, experimental; Contr., control. Number of presentations 6 7 8 10 6 8 7 9 9 7 7.7 (1.3) Follow-up (stimuli) Exp. Contr. 5 6 12 6 12 10 8 9 7 10 8.5 (2.5) 2 1 4 0 3 2 3 2 2 3 2.2 (1.1) 50 BASSO ET AL. TABLE 2 Reading Aloud Method: Number of Stimuli (Experimental and Control) Correctly Named by the Subjects at Baseline and Follow-Up, and Number of Presentations Necessary to Learn the 30 Experimental Words Baseline (stimuli) Subjects CA CA MDL MI NL MG FC LB FL VP Mean (SD) Exp. Contr. Number of presentations 3 1 3 6 3 5 3 2 3 0 2.9 (1.7) 4 2 3 3 1 1 2 2 2 0 2 (1.1) 6 9 7 7 8 6 8 7 8 7 7.3 (0.9) Follow-up (stimuli) Exp. Contr. 10 6 10 7 8 13 8 8 7 16 9.3 (3) 3 0 2 2 2 1 1 2 3 1 1.7 (0.9) Note. Exp., experimental: Contr., control. words with the repetition (7.7) or the reading aloud method (7.3) (Duncan’s test: p ⬍ .01). The repetition and the reading aloud methods did not differ from each other (Duncan’s test: not significant). Number of experimental words correctly remembered at follow-up. For each subject in each experimental group, the number of experimental and control pictures correctly named at baseline and follow-up was computed. The data were analyzed with an Anova with one within-subject factor (cueing method: repetition, reading aloud, and orthographic cue) and two between-subject factors (time, baseline vs follow-up; number of pictures, experimental vs control). Results showed three significant main effects: method (F ⫽ 10.168, df 2, 27, p ⫽ .0005), time (F ⫽ 148.654, df 1, 27, p ⫽ .0000), and number of pictures (F ⫽ 215.586, df 1, 27, p ⫽ .0000). TABLE 3 Orthographic Cueing Method: Number of Stimuli (Experimental and Control) Correctly Named by the Subjects at Baseline and Follow-Up, and Number of Presentations Necessary to Learn the 30 Experimental Words Baseline (stimuli) Subjects AG BG BL AG LC MP AP AG MP SF Mean (SD) Exp. Contr. 3 3 3 4 3 3 3 2 2 4 3.0 (0.7) 2 3 2 4 4 3 4 3 2 1 2.8 (1.0) Exp. ⫽ experimental, Contr. ⫽ control Number of presentations 5 5 6 5 5 5 5 5 6 4 5.1 (0.6) Follow-up (stimuli) Exp. Contr. 17 10 16 17 16 15 18 18 15 22 16.4 (3.0) 2 3 1 3 3 1 2 2 2 1 2.0 (0.8) ACQUISITION OF NEW WORDS 51 FIG. 2. Mean number of experimental stimuli correctly named by the three groups of control subjects at baseline and follow-up. The two-way interactions were also significant: cueing method ⫻ time (F ⫽ 20.420, df 2, 27, p ⫽ .0000), cueing method ⫻ number of pictures (F ⫽ 13.484, df 2, 27, p ⫽ .0000), and time ⫻ numbers of pictures (F ⫽ 260.943, df 2, 27, p ⫽ .0000). Finally, there was a significant three-way interaction: F ⫽ 21.407, df 2, 27, p ⫽ .0000. The interactions were further submitted to a planned comparison analysis. The results showed that for all methods the number of experimental pictures correctly named at follow-up was significantly higher than that at baseline (orthographic cue, F ⫽ 198.36, df 1, 27, p ⫽ .0000; reading, F ⫽ 45.25, df 1, 27, p ⫽ .0000; repetition, F ⫽ 25, 453, df1, 27, p ⫽ .0000). In addition, the mean number of experimental pictures correctly named with the orthographic cue at follow-up (16.4) was significantly higher than that with the repetition (8.5; F ⫽ 40.85, df 1, 27, p ⫽ .0000) and the reading aloud method (9.3; F ⫽ 27.06, df 1, 27, p ⫽ .0000), which did not differ from each other (F ⫽ 1.41, df 1, 27, p ⫽ .24). For the control pictures, with the orthographic and the repetition methods, the mean number of pictures correctly named at follow-up was significantly lower than that correctly named at baseline (orthographic cue, F ⫽ 198.36, df 1, 27 p ⫽ .02; repetition, F ⫽ 17.58, df 1, 27 p ⫽ .0000). The difference was not significant for the reading aloud method (F ⫽ 0.81, df 1, 27, p ⫽ .37). The graph in Fig. 3 illustrates these results. APHASIC PATIENTS RF is a 66-year-old right-handed university teacher with a degree in pedagogy. He suffered an occlusion of the left posterior cerebral artery with aphasia and right hemiparesis on January 1999. An MRI showed the presence of a large cortico-subcortical occipital-temporal lesion involving the thalamus and the internal and external capsule. In February 1999, his speech was fluent and grammatically correct with frequent word-finding difficulties. Comprehension was relatively spared (23/36 on the Token test; cutoff score, 29; De Renzi & Faglioni, 1978). Reading was impaired and reading treatment was started. In July he was reassessed; his comprehension was 52 BASSO ET AL. FIG. 3. Mean number of control stimuli correctly named by the three groups of control subjects at baseline and follow-up. within normal limits (Token test: 29/36) and his reading was much improved; wordfinding difficulties were still evident and more marked for nouns (11/30 correct) than for verbs (21/28 correct). All errors were omissions. The decision was made to submit the patient to the three experimental methods for the rehabilitation of anomia. MR is a 42-year-old right-handed man with a degree in law, who worked as an Italian Red Cross army officer. In December 1996 he underwent an evacuation of a haemorrhage due to the rupture of an artero-venous malformation of the posterior communicating artery. A CT-scan performed in January 1997 showed a hypodense area in the left temporal region. He then showed a full-blown Wernicke aphasia with severely impaired comprehension, fluent speech with semantic and phonemic paraphasias and word-finding difficulties. A rehabilitation program was started and when reassessed in October 1999 MR showed marked improvement. Comprehension was fair (Token test: 25/36), speech was fluent with word-finding difficulties more marked for nouns (11/30 correct) than for verbs (22/28 correct). His errors were mainly failures to respond. He was then submitted to the three experimental methods for the rehabilitation of anomia. Procedure Two hundred fourteen pictures of different semantic categories were selected; frequency varied from low to high. The pictures were presented once to the patients for three consecutive days, and the examiner waited 5 s for the response. The pictures the patients could not name and for which they always produced an omission were selected (103 for patient RF and 88 for patient MR). The selected pictures were then presented three times for comprehension. They were presented once with the correct name, once with a semantic alternative and once with an unrelated name (for table, for instance, they were told once table, once writing desk, and once car); the patient had to say whether the name was correct or not. Only the pictures for which the patients’ answers were always correct were chosen for the experimental naming therapy (92 for patient RF and 88 for patient MR). Orthographic Repetition Reading Methods 2 3 2 Exp Baseline 4 4 3 Contr 21 10 13 Exp 5 5 4 Contr Fifth day 16 8 5 Exp 5 4 5 Contr Follow-up 1 10 7 6 Exp 3 3 5 Contr Follow-up 2 TABLE 4 Number of Stimuli Correctly Named by RF, at Baseline, at the End of Treatment (Fifth Day), and at Follow-Ups 1 and 2 with Each Learning Method ACQUISITION OF NEW WORDS 53 54 BASSO ET AL. Therapy The selected pictures were subdivided in four subgroups of 23 (for RF) and 22 (for MR) pictures each, controlled for frequency of use. Three groups were used for the three learning methods and one as a control. The same control pictures were used for all therapy programs, which were performed on three subsequent weeks. For patient RF the order of presentation was orthographic cue, repetition, and reading aloud. For patient MR the order was repetition, orthographic cue, and reading. Before starting therapy the patients were presented with the control pictures and the group of pictures for the naming program to be started (baseline). The treatment was the same as for the control subjects with two differences. In the orthographic cueing method, the examiner waited 5 s (and not 3) before adding another letter to the preceding ones, and for all three methods the procedure was repeated three times a day for 5 consecutive days. Follow-up. At follow-up, performed independently for each therapy program, the patients were asked to name the control pictures and the pictures belonging to the last program performed. Patient RF was tested 1 and 12 weeks after the end of each treatment, and patient MR after 1, 3, and 5 weeks after the end of each treatment. RESULTS RF. Table 4 and Fig. 4 show, for each method, the number of pictures (experimental and control) correctly named during the baseline condition, at the end of treatment, after 1 (follow-up 1) and 12 weeks (follow-up 2). Naming accuracy at the end of treatment was significantly different from baseline for the words treated with the orthographic cueing and the reading methods: orthographic cueing, χ 2 17, p ⬍ .001; reading, χ 2 7.7, p ⬍ .01 (McNemar’s test). The difference was not significant for the repetition method. A second analysis compared the pictures correctly named during baseline with the pictures named at follow-ups 1 and 2. Naming accuracy was significantly different for the orthographic cueing method both at follow-up 1 (χ 2 10.5, p ⬍ .01) and at follow-up 2 (χ 2 4.9, p ⬍ .05; McNemar’s test). The differences for the reading and the repetition methods were not significant. FIG. 4. Number of experimental stimuli correctly named by RF at baseline, end of treatment (fifth day), and at follow-ups 1 and 2 with each learning method. Repetition Orthographic Reading Methods 4 2 2 Exp 2 5 4 Contr Baseline 17 22 16 Exp 2 4 4 Contr Fifth day 10 18 8 Exp 4 4 4 Contr Follow-up 1 8 18 6 Exp 3 4 3 Contr Follow-up 2 8 15 7 Exp 5 6 5 Contr Follow-up 3 TABLE 5 Number of Stimuli Correctly Named by MR at Baseline, End of Treatment (Fifth Day), and at Follow-Ups 1, 2, and 3 with Each Learning Method ACQUISITION OF NEW WORDS 55 56 BASSO ET AL. FIG. 5. Number of experimental stimuli correctly named by MR, at baseline, end of treatment (fifth day), and at follow-ups 1, 2, and 3 with each learning method. MR. Table 5 and Fig. 5 show, for each method, the number of pictures (experimental and control) correctly named during the baseline condition, at the end of treatment, after 1 (follow-up 1), 3 (follow-up 2), and 5 weeks (follow-up 3) after the end of treatment. Naming accuracy at the end of treatment was significantly different from baseline for all three methods (orthographic cueing, χ 2 18.05, p ⬍ .001; reading, χ 2 12, p ⬍ .001; repetition, χ 2 11, p ⬍ .001; McNemar’s test). A second analysis revealed that naming accuracy was still significantly different from baseline only for the orthographic cueing at follow-up 1 (χ 2 12.5, p ⬍ .001), follow-up 2 (χ 2 14, p ⬍ .001), and follow-up 3 (χ 2 11.07, p ⬍ .001; McNemar’s test). The differences for the reading and the repetition methods were not significant. DISCUSSION Not unexpectedly all the control subjects reached criterion with the three learning methods and retained some learning at follow-up. The use of the orthographic cue was found to be significantly more successful with regards both to the number of presentations necessary to reach criterion (which was lower for the orthographic method than for the other two methods), and to the number of ‘‘words’’ remembered at follow-up, a week later (which was higher for the orthographic method than for the other two methods). The other two methods did not differ. Results were similar for the two aphasic patients. Some learning was possible with all three methods (the difference was not significant with the repetition method for RF) but the patients did not learn all the experimental words. In addition, both RF and MR retained some learning at follow-ups, and this was significantly higher than baseline when words had been acquired with the orthographic cueing method. What aspects of the orthographic cueing method and of the other two methods, reading aloud and repetition, can explain their different effectiveness in learning? In all three cases, the target is activation of the phonological representation in the output lexicon. The search for the correct response, however, seems to be more under intentional control in the case of the orthographic cueing method than in reading or repetition; in these cases the response is given to the subject and all he or she has to do ACQUISITION OF NEW WORDS 57 is to transcode it from orthography to phonology in reading, and from input to output phonology in repetition. In normal subjects, active participation in the learning process is known to produce better retention than passive observation. The generation effect is the advantage in memory of self-produced as opposed to externally presented information (Slamecka & Graf, 1978). Researchers do not agree on the possible explanations of the generation effect. The various interpretations can be divided into two categories, those that implicate semantic memory explaining the effect as due to some more general principle of memory (such as depth of encoding, for instance) and those that attribute it to the intrinsic characteristics of the generation task (for a review, see Mc Elroy & Slamecka, 1982). Be it as it may, the important point for us here is that the generation effect has been proved sound. It has been found in a wide variety of tasks, such as frequency judgments (Green, 1988; Smith, 1996), procedural learning (Vakil, Hoffman, & Myzliek, 1998), and acquisition of new words (Mc Elroy & Slamecka, 1982). Our results with normal control subjects confirm the effectiveness of generation over less effortful methods (reading and repetition) although other explanations cannot be ruled out. An obvious difference between the three methods lies in the amount of time allowed for each response, which is longer in the orthographic cueing method. However, in a pilot study with control subjects carried out in order to identify the number of new ‘‘words’’ needed to be learned, leaving more time for the answer did not help the subjects. For aphasic patients, however, time may be an important factor. The following step consisted in evaluating the three methods with aphasic subjects. We argued that temporal absence of a phonological representation in the mother tongue or absence of a phonological representation in a second language in normal controls can be compared to functional damage to the phonological output lexicon in aphasic patients and that the same learning strategies could be successful, albeit less effective in aphasic patients because of a general loss of capacity to learn (Tikofsky, 1971; Carson et al., 1968). The effect of the three methods would be less but the orthographic cueing method, being the most effective one in normal controls, would still induce some learning. Our results confirm this hypothesis. To conclude, let us briefly recapitulate our main points. We argued that as long as it has not proven to be ineffective, aphasia therapy should be directed at the functional damage and not to finding alternative solutions, and that aphasic patients with specific functional damages can be compared to normal controls in special situations. We also argued that strategies that work with normal subjects in such situations can also be successful with aphasic patients. We therefore identified a highly effective learning method with normal subjects and verified its effectiveness with aphasic patients. 58 BASSO ET AL. Appendix Fam M H L H H H M H H L L M L L M M L M H L M H M L L L H M L H H L M M L L M L H L L L M H M M H M H H H H L L M M L H L H Pictures Ago (needle) Camicia (shirt) Cigno (swan) Telefono (telephone) Libro (book) Cucina (kitchen) Pesce (fish) Naso (nose) Bicchiere (glass) Capra (goat) Leone (lion) Zucca (pumpkin) Arpa (harp) Corona (crown) Arancia (orange) Stella (star) Struzzo (ostrich) Sedano (celery) Porta (door) Bruco (caterpillar) Freccia (arrow) Gamba (leg) Lucchetto (padlock) Canguro (kangaroo) Giraffa (giraffe) Foca (seal) Pantaloni (trousers) Scala (ladder) Lumaca (snail) Calzino (sock) Vestito (dress) Serpente (snail) Carota (carrot) Cappello (hat) Mulino (mill) Pecora (sheep) Pinze (pliers) Scimmia (monkey) Stampella (hanger) Orso (bear) Ancora (anchor) Topo (mouse) Ciliegia (cherry) Occhio (eye) Palla (ball) Chiodo (nail) Chiave (key) Sega (saw) Braccio (arm) Dito (finger) Frigorifero (refrigerator) Cane (dog) Leopardo (leopard) Botte (barrel) Anguria (watermelon) Lima (file) Telaio (loom) Letto (bed) Scarafaggio (beetle) Capelli (hair) Note. Fam, familiarity; H, high; M, medium; L, low. Invented words Lasba Grole Zaclo Nuspo Norli Velba Mible Bippo Vorpa Pilca Nunco Lansi Cirli Bepri Pirga Lorba Bleti Lutre Fimpo Dresi Svife Dongi Lerpo Milvo Svuna Gilne Tucce Revra Filco Tiblo Rundo Galpo Silpo Lecri Gluve Lorfe Pasbi Clesi Tabri Drelo Fitre Trulo Nirgo Senci Relga Sdeta Sbulo Lovri Dippa Zirve Zilci Zunde Nipra Pruso Nulge Rembe Docle Tepro Furra Dunlo ACQUISITION OF NEW WORDS 59 REFERENCES Behrmann, M. (1987). The rites of righting writing. Cognitive Neuropsychology, 4, 365–384. Byng, S. (1988). Sentence processing deficits: Theory and therapy. Cognitive Neuropsychology, 5, 629– 676. Caramazza, A. (1989). Cognitive rehabilitation: An unfulfilled promise? In X. Seron and G. Deloche (Eds.), Cognitive approaches to neuropsychological rehabilitation. Hillsdale, NJ: Erlbaum. Caramazza, A., & Hillis, A. E. (1990). Where do semantic errors come from? Cortex, 26, 95–122. Carson, D. H., Carson, F. E., & Tikofsy, R. S. (1968). On learning characteristics of the adult aphasics. Cortex, 4, 92–112. De Renzi, E., & Faglioni, P. (1978). Normative data and the screening power of a shortened version of the Token test. Cortex, 14, 41–49. Freud, S. (E. Stengel, Trans.) (1953). On aphasia New York: International University Press. (Original work published 1891) Green, R. L. (1988). Generation effects in frequency judgements. Journal of Experimental Psychology: Learning, Memory and Cognition, 14, 298–304. Hillis, A. E. (1989). Efficacy and generalization of treatment for aphasic naming errors. Archives of Physical Medicine and Rehabilitation, 70, 632–636. Hillis, A. E., & Caramazza, A. (1987). Model-driven treatment of dysgraphia. In R.H. Brookshire (Ed.), Clinical aphasiology (pp. 84–105). Minneapolis: BRK Publishers. Howard, D., & Hatfield (1987). Aphasia therapy. Historical and contemporary issues. Hove, UK; LEA. Howard, D., & Horchard-Lisle, V.M. (1984). On the origin of semantic errors in naming: Evidence from the case of a global aphasic. Cognitive Neuropsychology, 1, 163–190. Howard, D., Patterson, K. E., Franklin, S., Orchard-Lisle, V. M., & Morton, J. (1985a.) The facilitation of picture naming in aphasia. Cognitive Neuropsychology, 2, 49–80. Howard, D., Patterson, K. E., Franklin, S., Orchard-Lisle V. M., & Morton, J. (1985b.) The treatment of word retrieval deficits in aphasia: A comparison of two therapy methods. Brain, 108, 817–829. Marshall, J. C., Pound, C., White-Thompson M., & Pring, T. (1990). The use of picture/matching tasks to assist word retrieval in aphasic patients. Aphasiology, 4, 167–184. Mc Elroy, L. A., & Slamecka, N. J. (1982). Memorial consequences of generating nonwords: Implications for semantic-memory interpretations of the generation effect. Journal of Verbal Learning and Verbal Behavior, 21, 249–259. Miceli, G., Amitrano, A., Capasso, R. & Caramazza, A. (1996). The treatment of anomia resulting from output lexical damage: Analysis of two cases. Brain and Language, 52, 150–174. Miceli, G., Giustolisi, L., & Caramazza, A. (1991). The interaction of lexical and nonlexical processing mechanisms: Evidence from anomia. Cortex, 27, 57–80. Myers-Pease, D., & Goodglass, H. (1978). The effects of cuing on picture naming in aphasia. Cortex, 14, 178–189. Slameka, N.J., & Graf, P. (1978). The generation effect: Delineation of a phenomenon. Journal of Experimental Psychology: Human Learning and Memory, 4, 592–604. Smith, M. L. (1996). Recall of frequency of occurrence of self-generated and examiner-provided words after frontal or temporal lobectomy. Neuropsychologia, 34, 553–563. Snodgrass, J. G., & Vanderwart, M. (1980). A standardized set of 260 pictures: Norms for name agreement, image agreement, familiarity, and visual complexity. Journal of Experimental Psychology: Human Learning and Memory, 6, 174–215. Tikofsky, R. S. (1971). Two studies of verbal learning by adult aphasics. Cortex, 7, 106–125. Vakil, E., Hoffman, Y., & Myzliek, D. (1998). Active versus passive procedural learning in older and younger adults. Neuropsychological Rehabilitation, 8, 31–41.