Object Naming Ability of Adults with Written Language Difficulties Alisa Cantwell Inverness District School Board Port Hastings, Nova Scotia
Hyla Rubin University of Toronto Toronto, Ontario
Several studies have observed that school-aged, reading-disabled children have object-naming problems. In addition, significant positive relationships between object-naming ability and reading and spelling skills have been observed for this population. The co-occurrence of these problems has been explained by common underlying phonological deficiencies. Because written language problems can persist beyond the school-aged years, the purpose of this study was to examine object-naming ability and the relationship between object naming and written language of adults. Twenty-two adults, half with written language difficulties and half without, performed four tasks: object naming, object recognition, reading, and spelling. Significant positive relationships were obtained between object-naming ability and reading ability, object-naming ability and spelling This research was funded by Grant A2008 from the Natural Sciences and Engineering Research Council of Canada to the second author. We would like to thank the adults from the Continuing Education Department of Mohawk College of Applied Arts and Technolog~ Hamilton, Ontario, and from the Hamilton Public Library Literacy Program, for their willingness to participate in this stud)~ Requests for reprints should be addressed to: Hyla Rubin, Haskins Laboratories, 270 Crown St., New Haven, Ct., 06511. Annals of Dyslexia, Vol. 42, 1992. Copyright © 1992 by The Orton Dyslexia Society ISSN 0736-9387
179
180
OmGZNS, ParrEaNS, ANO PROGNOSZS perfvrmance, and reading and spelling performance. In addition to phonological deficiencies, the results indicated that adult poor readers and spellers lack knowledge of the orthographic structure of words. These findings suggest that problems underlying object naming and written language do not resolve with cognitive maturation or additional years of experience with language and should be addressed in the early school-aged years.
Introduction
Object-naming problems are characteristic of children with written language difficulties (Denckla and Rudel 1976; German 1979; Katz 1986; Rubin and Liberman 1983; Wolf 1982). In fact, object-naming performance in young prereaders has been found to predict later written language ability (Jansky and deHirsh 1972; Wolf and Goodglass 1986). To date, however, all studies examining the relationship between object naming and written language have involved school-aged children. Because written language problems (i.e., difficulties with reading, spelling, and written expression) can persist beyond the school-aged years, the present study was designed to determine whether objectnaming problems also continue into adulthood and to examine whether a relationship between object-naming and written language difficulties exists for this population. In order to name an object, a number of processes must occur (Goodglass 1980; Wolf 1982). Using perceptual and semantic information to recognize an object, the phonological representation of its name must be located in long-term memory. Once located, this representation must be processed and stored in short-term memory while production of the name is being planned and completed. Although difficulty at any of these stages could result in naming failure, recent findings suggest that naming errors primarily reflect phonological deficiencies in the naming process. Specifically, naming problems could occur when the phonological representation of an object name is incomplete, or located incorrectly, or when there is a deficiency in the processing of the phonological representation after it has been accessed (Katz 1986). Analysis of erroneous responses produced by reading-disabled children reveals that more errors are semantically related than phonologicaUy related to the intended words (Katz 1986; Rubin and Liberman 1983; Wolf 1982). It had been assumed until recently that the preponderance of semantic errors reflected semantic deficiencies in the naming process. Current studies have shown, however, that children who are poor readers not only have receptive knowledge of object names they are unable to produce (Katz 1986; Rubin and Liberman 1983; Rubin, Bernstein, and Katz 1989; Rubin, Zimmerman, and Katz
OBIECT-NAMINGABxrrrY oF ADULTS
181
1989; Wolf and Goodglass 1986), but they are also able to provide semantic information when asked about the objects they cannot name (Katz 1986). Since it has been demonstrated that the semantic information about an object name is available to poor readers, an alternative explanation proposes that semantically related naming errors may reflect phonological deficiencies, either in being unable to access the correct phonological representation, early in the retrieval process (Rubin and Liberman 1983), or in processing the phonological representation once it is accessed (Katz 1986). In contrast, those object-naming errors that bear a phonetic resemblance to the target name (e.g., acorn/unicorn) suggest that the correct name either was located but was processed inefficiently in short-term memory, or that a phonetically related object name was accessed instead. In either case, some degree of sensitivity to the phonological structure of the intended word has been demonstrated when this type of error is considered in relation to errors that are semantically, but not phonetically, related to the intended word (e.g., horse/ unicorn). Phonetically related errors tend to have the same number of syllables and the same initial phoneme as the target word (Katz 1986; Rubin and Liberman 1983). Furthermore, both frequency and length of the name have been found to influence naming performance (Katz 1986). Low-frequency words are named incorrectly more often than high-frequency words and more three- and four-syllable words are named incorrectly than one- and two-syllable object names. Katz (1986) suggested that frequency of a word would affect naming ability because low-frequency words, which are encountered infrequently, may not have as complete a representation as high-frequency words. Length of a word would influence naming ability because as the length of a word increases, the amount of phonological information to be stored and processed also increases. Studies examining the types of cues that improve naming ability also support the hypothesis that object-naming errors result from phonological deficiencies. Semantic cues, including superordinate categories, functions, and locations, have been found to be ineffective in improving the naming performance of reading-disabled children (Rubin, Bernstein, and Katz 1989), whereas initial phoneme cues (Rubin, Bernstein, and Katz 1989; Rubin, Zimmerman, and Katz 1989) and rhyming word cues (Rubin, Zimmerman, and Katz 1989) have been found to be highly effective in increasing naming success. This evidence, in conjunction with findings already reviewed, suggests that poor readers already have the semantic information available to them for naming, but do not demonstrate the phonological skill needed to either access target words correctly or to preserve that information efficiently in short-term memory once it is accessed.
182
OmCzNS, PATTBPaVS,aND PROGNOSBS
Recent research has suggested that phonological deficiendes not only help to explain object-naming problems, but are also strongly associated with reading and spelling difficulties in many children (Katz 1986; Rubin, Zimmerman, and Katz 1989; see also discussions by Liberman and Shankweiler 1985; Stanovich 1985). Both reading and writing depend on accurate and complete phonological representations and effident processing of these representations (in addition to knowledge about the orthographic structure of words). In reading, the phonology that is reflected in the orthography of the printed word is used to locate the phonological representation in long-term memory. (See Seidenberg 1985 for a discussion of the specific time course of phonological code activation, which varies with reading ability.) In writing the name of an object, the way in which the representation is located is the same as in object naming. Perceptual and semantic information is used in these situations. Most writing tasks however do not require the production of the name of an object in view. More often a person is composing sentences and paragraphs based around a general theme. In these situations semantic thought processes are used to locate the phonological representations. Thus, the manner in which a phonological representation is located may differ for reading, writing, and object naming, but all tasks require a phonological representation, and phonological processes are required once the representation is located to hold the representation in short-term memory. This explanation can help to explain the co-occurrence of object-naming problems and written language difficulties--both can result from phonological deficiencies. A significant relationship between object-naming and reading abilities has been established in all of the studies to date, and Rubin and Liberman (1983) have examined the relationship between objectnaming performance and writing performance. In their study no relationship was found between correctly named and correctly spelled items. This finding is surprising given that poor readers are also poor spellers. However, their subjects were language-disabled children and the range of performance of these children on the spelling task was too restricted for a correlation to be detected. Although Rubin and Liberman (1983) did not find a relationship between accuracy of naming and accuracy of spelling, a significant positive relationship between retrieval of verbal labels on object naming and spelling tasks was obtained. The present study was conducted to obtain further evidence for the hypothesis that phonological deficiencies underlie both objectnaming problems and written-language difficulties by examining object-naming abilities of adult poor readers and spellers. Although written-language problems can persist into adulthood (Frauenheim and Heckerl 1983; Johnson 1980), the object-naming abilities of adults
OmlEC"r-N.4MXNG ABXLn'Y OF ADVLTS
183
with written-language problems have not been examined. Only Stirling and Miles (1988) tested object-naming abilities beyond elementary school years. In their study, skills of dyslexic boys between the ages of 11 and 18 were examined. No evidence of object-naming problems was obtained for this population. However, their stimulus set was very small (n = 12) and it contained relatively commonly-occurring objects (for example, "strings" of a racket and "eye" of a needle). Thus, Stirling and Miles (1988)may not have employed a sensitive enough measure to detect object-naming problems in this population. An alternative hypothesis could be proposed to explain the absence of object-naming problems in the dyslexic adolescents studied by Stifling and Miles (1988). That is, cognitive maturation or additional years of exposure to language may resolve object-naming difficulties. If object-naming problems result from phonological deficiencies, as suggested from research on children, then phonological deficiencies should not be observed in adults with written-language problems. However, several studies have found that adults with written-language difficulties lack awareness of the phonological structure of spoken language. These adults perform poorly on phoneme segmentation tasks (Byrne and Ledez 1983; Liberman et al. 1985; Morais et al. 1979; Pratt and Brady 1988; Read and Ruyter 1985), phoneme addition tasks (Morais et al. 1979; Read and Ruyter 1985), and word-length judgment tasks (Pratt and Brady 1988). Thus, adults with written-language difficulties appear to have phonological deficiencies which adversely affect performance on phonological-awareness tasks. These phonological deficiencies may also result in object-naming problems. In fact, children with reading disabilities have problems on both object-naming tasks and phonological-awareness tasks (Katz 1986; Rubin, Zimmerman, and Katz 1989; Rubin et al. 1991). Therefore, object-naming problems may also be characteristic of adults with written-language difficulties. To determine if object-naming problems persist into adulthood, the present study compared the performance of adults with writtenlanguage problems to the performance of a control group of adults without written-language problems on an object-naming task. If adults with written-language difficulties have object-naming problems, fewer objects should be named by this group than the control group. In order to separate object-naming difficulties from a general vocabulary deficit, performance on the object-naming task was compared to performance on an object-recognition task. If adults are unable to recognize objects they cannot name, a general vocabulary deftcit exists. On the other hand, if adults are unable to name objects that they can recognize, an object-naming problem exists. To examine whether a relationship exists between object-naming performance and written-language skills for this population, all sub-
184
O~aINS, P,4vrEI~cs, ~ o PROaNOSES
jects also performed a reading task and a spelling task. The same object names were used in all tasks. Previous research has examined the relationship between object-naming and written-language difficulties by correlating object-naming performance on a set of items with performance on standardized reading tests (Katz 1986; Rubin and Liberman 1983; Wolf 1982; Wolf and Goodglass 1986) or with writing performance on a subset of the items (Rubin and Liberman 1983). Examining performance on the same items across tasks may provide a more sensitive measure of determining whether a relationship exists between object-naming and written-language difficulties. Furthermore, since previous research has shown that adults with written-language difficulties not only have phonological deficiencies but also lack orthographic knowledge (Liberman et al. 1985; Read and Ruyter 1985), we wanted to compare performance on the oral and written language tasks. We reasoned that if only phonological deficiencies contribute to written-language problems, then performance on the object-naming task and written-language tasks should be similar. However, if the adults with written-language problems are unable to read and write the names of objects they could name, then evidence of lack of orthographic knowledge would be obtained.
Method Subjects Experimental Group. Eleven adult volunteers with written-language problems formed this group. The group consisted of nine males and two females. Ages of the adults ranged from 22 to 62 years, with a mean of 29 years. Highest education level attained by these adults ranged from grade 2 to grade 12 plus vocational training. Three adults repeated at least one grade in school. Only one adult received extra help in school (tutoring). Four of the adults were employed at time of testing (two laborers, one housekeeper, one carpenter). All were monolingual speakers of English, with no evidence of sensory or intellectual deficiencies. Ten adults were attending either an Adult Literacy Program or a Job Readiness Program in which they were receiving help for their reading and writing problems. The adult not attending either of these programs was contacted personally. Three adults reported that they were "late talkers." However, none of the adults in this group received speech/language services besides those offered in the literacy programs in which they were enrolled. To examine whether the adults in the experimental group had orallanguage difficulties in addition to written-language problems, the Producing Formulated Sentences subtest of the Clinical Evaluation of
OBJECT-NAMINGABzLr~ oF AOUrTS
185
Language Functions test (Semel and Wiig 1980) was administered. In this subtest the subject is given a word and asked to produce orally a sentence using this word. A total of 12 words was given. Sentences were scored as either grammatically correct or incorrect. Percentage of grammatically correct sentences ranged for 25 percent to 100 percent, with a mean of 72 percent. Thus, using this rough estimation of oral language skills, it appears that in addition to written-language problems, these subjects also have oral-language difficulties. However, most of the grammatically incorrect sentences consisted of colloquialtype errors (for example, "Me and my w i f e . . . ") or incomplete sentences (for example, "If you're g o o d . . . "). Control Group. Eleven adults without written-language problems formed the Control Group. There were five males and six females, ranging in age from 19 to 52 years, with a mean age of 28 years. Highest level of schooling achieved ranged from grade 7 to grade 12 plus vocational training. None of the subjects in this group repeated any grades or received any extra help during school. All but two of the adults were employed at time of testing (three laborers, four waiter/waitresses, two certified nursing assistants). All were monolingual speakers of English. None of the subjects appeared to have any sensory or intellectual deficiencies. None of the subjects reported a history of speech/ language problems and no subject received any speech/language services. The mean score on the Producing Formulated Sentences subtest of the Clinical Evaluation of Language Functions test was 94 percent (ranging from 83 percent to 100 percent). Most of the errors made by these adults consisted of incomplete sentences (for example, "Only if you are in my s h o e s . . . "). Materials
Forty-eight object names were selected based on the following characteristics: 1) frequency of name (high vs. low); and 2) length of name (short vs. long). Frequency of the name of each object was determined using Carroll, Davies, and Richman's (1971) word frequency count. The name of an object was classified as "High" if it occurred at a frequency rate of at least one per million and "Low" if it occurred at a frequency rate of less than one per million. Length was defined using the number of syllables in the name of an object. Objects with one- or two-syllable names were said to be "Short" names whereas three- or four-syllable names were said to be "Long" names. Using these two characteristics, four groups of object names were created (with mean frequency of occurrence given here): 1) High frequency/Short names (5.275); 2) High frequency/Long names (5.267); 3) Low frequency/ Short names (0.419); and, 4) Low frequency/Long names (0.440). Each group contained twelve object names. No significant differences ex-
186
OmGi1vs, PAvrE~avS,ANY PRoa~voses
isted in the frequency of occurrence between "Short" and "Long" object names for either "Low" frequency, F(1,23) = 0.045, or "High" frequency, F(1,23) = 0.000, object names. All 48 stimuli were used in each of the four tasks in the study (see Appendix). For three of the tasks (object recognition, object naming, and spelling) a photograph was taken of each object. For the objectrecognition task, photographs of the twelve objects in each group were divided into three sets of four photographs. This procedure resulted in a total of twelve sets of photographs. Photographs from each set were mounted on paper in a 2 × 2 arrangement. For the fourth task, the reading task, the name of each object was printed on a separate 3" × 5" index card.
Procedure Each adult was tested individually in one session which ranged from 30 to 60 minutes. Each adult performed four tasks in the following order: 1) spelling task; 2) object-naming task; 3) reading task; and, 4) object-recognition task. For each task the stimuli were presented in a random order, but the presentation order for each task remained fixed across subjects. Spelling Task. Photographs of objects were presented one at a time. The subject was asked to print the name of the object in the photograph on a sheet of paper containing blanks numbered 1 to 48. Object-Naming Task. The procedure for this task was similar to the spelling task except the subject was asked to name the object in the photograph. When an incorrect response was given, the subject was asked to provide another name for the object. The subject was allowed three trials to name an object correctly. The experimenter recorded all responses. Reading Task. In this task the subject was presented with index cards one at a time and asked to read the word printed on the card. The experimenter recorded all responses. Object-Recognition Task. The subject was shown a set of four photographs. The experimenter named one of the objects. The subject was asked to point to the object named by the experimenter.
Results Object Naming To determine if adults with written-language problems also experience object-naming difficulties, performance on the object-recognition and object-naming tasks was compared for each group. For each subject, any object not recognized was eliminated from the object-
OBJECT-NAMINGASXLXTYOr ADULTS
187
naming calculations to ensure that any differences between the groups could not be attributed to a vocabulary deficit. For the experimental group a mean number of two objects per subject was eliminated. For the control group, no objects were eliminated from the analysis because all subjects recognized all the objects. Scores were converted to percentage correct. In the object-naming task subjects were allowed three trials to name an object accurately. The mean number of objects recognized and objects n a m e d correctly on the first trial by each group is presented in Table I. A 2 x 2 split-plot ANOVA was conducted, involving Group (experimental, control) by Task (object recognition, object naming) with repeated measures on the last factor. The analysis yielded a significant main effect of Group, F(1,20) = 18.698, p = .0006, and of Task, F(1,20) = 145.637, p
Hyla Rubin University of Toronto Toronto, Ontario
Several studies have observed that school-aged, reading-disabled children have object-naming problems. In addition, significant positive relationships between object-naming ability and reading and spelling skills have been observed for this population. The co-occurrence of these problems has been explained by common underlying phonological deficiencies. Because written language problems can persist beyond the school-aged years, the purpose of this study was to examine object-naming ability and the relationship between object naming and written language of adults. Twenty-two adults, half with written language difficulties and half without, performed four tasks: object naming, object recognition, reading, and spelling. Significant positive relationships were obtained between object-naming ability and reading ability, object-naming ability and spelling This research was funded by Grant A2008 from the Natural Sciences and Engineering Research Council of Canada to the second author. We would like to thank the adults from the Continuing Education Department of Mohawk College of Applied Arts and Technolog~ Hamilton, Ontario, and from the Hamilton Public Library Literacy Program, for their willingness to participate in this stud)~ Requests for reprints should be addressed to: Hyla Rubin, Haskins Laboratories, 270 Crown St., New Haven, Ct., 06511. Annals of Dyslexia, Vol. 42, 1992. Copyright © 1992 by The Orton Dyslexia Society ISSN 0736-9387
179
180
OmGZNS, ParrEaNS, ANO PROGNOSZS perfvrmance, and reading and spelling performance. In addition to phonological deficiencies, the results indicated that adult poor readers and spellers lack knowledge of the orthographic structure of words. These findings suggest that problems underlying object naming and written language do not resolve with cognitive maturation or additional years of experience with language and should be addressed in the early school-aged years.
Introduction
Object-naming problems are characteristic of children with written language difficulties (Denckla and Rudel 1976; German 1979; Katz 1986; Rubin and Liberman 1983; Wolf 1982). In fact, object-naming performance in young prereaders has been found to predict later written language ability (Jansky and deHirsh 1972; Wolf and Goodglass 1986). To date, however, all studies examining the relationship between object naming and written language have involved school-aged children. Because written language problems (i.e., difficulties with reading, spelling, and written expression) can persist beyond the school-aged years, the present study was designed to determine whether objectnaming problems also continue into adulthood and to examine whether a relationship between object-naming and written language difficulties exists for this population. In order to name an object, a number of processes must occur (Goodglass 1980; Wolf 1982). Using perceptual and semantic information to recognize an object, the phonological representation of its name must be located in long-term memory. Once located, this representation must be processed and stored in short-term memory while production of the name is being planned and completed. Although difficulty at any of these stages could result in naming failure, recent findings suggest that naming errors primarily reflect phonological deficiencies in the naming process. Specifically, naming problems could occur when the phonological representation of an object name is incomplete, or located incorrectly, or when there is a deficiency in the processing of the phonological representation after it has been accessed (Katz 1986). Analysis of erroneous responses produced by reading-disabled children reveals that more errors are semantically related than phonologicaUy related to the intended words (Katz 1986; Rubin and Liberman 1983; Wolf 1982). It had been assumed until recently that the preponderance of semantic errors reflected semantic deficiencies in the naming process. Current studies have shown, however, that children who are poor readers not only have receptive knowledge of object names they are unable to produce (Katz 1986; Rubin and Liberman 1983; Rubin, Bernstein, and Katz 1989; Rubin, Zimmerman, and Katz
OBIECT-NAMINGABxrrrY oF ADULTS
181
1989; Wolf and Goodglass 1986), but they are also able to provide semantic information when asked about the objects they cannot name (Katz 1986). Since it has been demonstrated that the semantic information about an object name is available to poor readers, an alternative explanation proposes that semantically related naming errors may reflect phonological deficiencies, either in being unable to access the correct phonological representation, early in the retrieval process (Rubin and Liberman 1983), or in processing the phonological representation once it is accessed (Katz 1986). In contrast, those object-naming errors that bear a phonetic resemblance to the target name (e.g., acorn/unicorn) suggest that the correct name either was located but was processed inefficiently in short-term memory, or that a phonetically related object name was accessed instead. In either case, some degree of sensitivity to the phonological structure of the intended word has been demonstrated when this type of error is considered in relation to errors that are semantically, but not phonetically, related to the intended word (e.g., horse/ unicorn). Phonetically related errors tend to have the same number of syllables and the same initial phoneme as the target word (Katz 1986; Rubin and Liberman 1983). Furthermore, both frequency and length of the name have been found to influence naming performance (Katz 1986). Low-frequency words are named incorrectly more often than high-frequency words and more three- and four-syllable words are named incorrectly than one- and two-syllable object names. Katz (1986) suggested that frequency of a word would affect naming ability because low-frequency words, which are encountered infrequently, may not have as complete a representation as high-frequency words. Length of a word would influence naming ability because as the length of a word increases, the amount of phonological information to be stored and processed also increases. Studies examining the types of cues that improve naming ability also support the hypothesis that object-naming errors result from phonological deficiencies. Semantic cues, including superordinate categories, functions, and locations, have been found to be ineffective in improving the naming performance of reading-disabled children (Rubin, Bernstein, and Katz 1989), whereas initial phoneme cues (Rubin, Bernstein, and Katz 1989; Rubin, Zimmerman, and Katz 1989) and rhyming word cues (Rubin, Zimmerman, and Katz 1989) have been found to be highly effective in increasing naming success. This evidence, in conjunction with findings already reviewed, suggests that poor readers already have the semantic information available to them for naming, but do not demonstrate the phonological skill needed to either access target words correctly or to preserve that information efficiently in short-term memory once it is accessed.
182
OmCzNS, PATTBPaVS,aND PROGNOSBS
Recent research has suggested that phonological deficiendes not only help to explain object-naming problems, but are also strongly associated with reading and spelling difficulties in many children (Katz 1986; Rubin, Zimmerman, and Katz 1989; see also discussions by Liberman and Shankweiler 1985; Stanovich 1985). Both reading and writing depend on accurate and complete phonological representations and effident processing of these representations (in addition to knowledge about the orthographic structure of words). In reading, the phonology that is reflected in the orthography of the printed word is used to locate the phonological representation in long-term memory. (See Seidenberg 1985 for a discussion of the specific time course of phonological code activation, which varies with reading ability.) In writing the name of an object, the way in which the representation is located is the same as in object naming. Perceptual and semantic information is used in these situations. Most writing tasks however do not require the production of the name of an object in view. More often a person is composing sentences and paragraphs based around a general theme. In these situations semantic thought processes are used to locate the phonological representations. Thus, the manner in which a phonological representation is located may differ for reading, writing, and object naming, but all tasks require a phonological representation, and phonological processes are required once the representation is located to hold the representation in short-term memory. This explanation can help to explain the co-occurrence of object-naming problems and written language difficulties--both can result from phonological deficiencies. A significant relationship between object-naming and reading abilities has been established in all of the studies to date, and Rubin and Liberman (1983) have examined the relationship between objectnaming performance and writing performance. In their study no relationship was found between correctly named and correctly spelled items. This finding is surprising given that poor readers are also poor spellers. However, their subjects were language-disabled children and the range of performance of these children on the spelling task was too restricted for a correlation to be detected. Although Rubin and Liberman (1983) did not find a relationship between accuracy of naming and accuracy of spelling, a significant positive relationship between retrieval of verbal labels on object naming and spelling tasks was obtained. The present study was conducted to obtain further evidence for the hypothesis that phonological deficiencies underlie both objectnaming problems and written-language difficulties by examining object-naming abilities of adult poor readers and spellers. Although written-language problems can persist into adulthood (Frauenheim and Heckerl 1983; Johnson 1980), the object-naming abilities of adults
OmlEC"r-N.4MXNG ABXLn'Y OF ADVLTS
183
with written-language problems have not been examined. Only Stirling and Miles (1988) tested object-naming abilities beyond elementary school years. In their study, skills of dyslexic boys between the ages of 11 and 18 were examined. No evidence of object-naming problems was obtained for this population. However, their stimulus set was very small (n = 12) and it contained relatively commonly-occurring objects (for example, "strings" of a racket and "eye" of a needle). Thus, Stirling and Miles (1988)may not have employed a sensitive enough measure to detect object-naming problems in this population. An alternative hypothesis could be proposed to explain the absence of object-naming problems in the dyslexic adolescents studied by Stifling and Miles (1988). That is, cognitive maturation or additional years of exposure to language may resolve object-naming difficulties. If object-naming problems result from phonological deficiencies, as suggested from research on children, then phonological deficiencies should not be observed in adults with written-language problems. However, several studies have found that adults with written-language difficulties lack awareness of the phonological structure of spoken language. These adults perform poorly on phoneme segmentation tasks (Byrne and Ledez 1983; Liberman et al. 1985; Morais et al. 1979; Pratt and Brady 1988; Read and Ruyter 1985), phoneme addition tasks (Morais et al. 1979; Read and Ruyter 1985), and word-length judgment tasks (Pratt and Brady 1988). Thus, adults with written-language difficulties appear to have phonological deficiencies which adversely affect performance on phonological-awareness tasks. These phonological deficiencies may also result in object-naming problems. In fact, children with reading disabilities have problems on both object-naming tasks and phonological-awareness tasks (Katz 1986; Rubin, Zimmerman, and Katz 1989; Rubin et al. 1991). Therefore, object-naming problems may also be characteristic of adults with written-language difficulties. To determine if object-naming problems persist into adulthood, the present study compared the performance of adults with writtenlanguage problems to the performance of a control group of adults without written-language problems on an object-naming task. If adults with written-language difficulties have object-naming problems, fewer objects should be named by this group than the control group. In order to separate object-naming difficulties from a general vocabulary deficit, performance on the object-naming task was compared to performance on an object-recognition task. If adults are unable to recognize objects they cannot name, a general vocabulary deftcit exists. On the other hand, if adults are unable to name objects that they can recognize, an object-naming problem exists. To examine whether a relationship exists between object-naming performance and written-language skills for this population, all sub-
184
O~aINS, P,4vrEI~cs, ~ o PROaNOSES
jects also performed a reading task and a spelling task. The same object names were used in all tasks. Previous research has examined the relationship between object-naming and written-language difficulties by correlating object-naming performance on a set of items with performance on standardized reading tests (Katz 1986; Rubin and Liberman 1983; Wolf 1982; Wolf and Goodglass 1986) or with writing performance on a subset of the items (Rubin and Liberman 1983). Examining performance on the same items across tasks may provide a more sensitive measure of determining whether a relationship exists between object-naming and written-language difficulties. Furthermore, since previous research has shown that adults with written-language difficulties not only have phonological deficiencies but also lack orthographic knowledge (Liberman et al. 1985; Read and Ruyter 1985), we wanted to compare performance on the oral and written language tasks. We reasoned that if only phonological deficiencies contribute to written-language problems, then performance on the object-naming task and written-language tasks should be similar. However, if the adults with written-language problems are unable to read and write the names of objects they could name, then evidence of lack of orthographic knowledge would be obtained.
Method Subjects Experimental Group. Eleven adult volunteers with written-language problems formed this group. The group consisted of nine males and two females. Ages of the adults ranged from 22 to 62 years, with a mean of 29 years. Highest education level attained by these adults ranged from grade 2 to grade 12 plus vocational training. Three adults repeated at least one grade in school. Only one adult received extra help in school (tutoring). Four of the adults were employed at time of testing (two laborers, one housekeeper, one carpenter). All were monolingual speakers of English, with no evidence of sensory or intellectual deficiencies. Ten adults were attending either an Adult Literacy Program or a Job Readiness Program in which they were receiving help for their reading and writing problems. The adult not attending either of these programs was contacted personally. Three adults reported that they were "late talkers." However, none of the adults in this group received speech/language services besides those offered in the literacy programs in which they were enrolled. To examine whether the adults in the experimental group had orallanguage difficulties in addition to written-language problems, the Producing Formulated Sentences subtest of the Clinical Evaluation of
OBJECT-NAMINGABzLr~ oF AOUrTS
185
Language Functions test (Semel and Wiig 1980) was administered. In this subtest the subject is given a word and asked to produce orally a sentence using this word. A total of 12 words was given. Sentences were scored as either grammatically correct or incorrect. Percentage of grammatically correct sentences ranged for 25 percent to 100 percent, with a mean of 72 percent. Thus, using this rough estimation of oral language skills, it appears that in addition to written-language problems, these subjects also have oral-language difficulties. However, most of the grammatically incorrect sentences consisted of colloquialtype errors (for example, "Me and my w i f e . . . ") or incomplete sentences (for example, "If you're g o o d . . . "). Control Group. Eleven adults without written-language problems formed the Control Group. There were five males and six females, ranging in age from 19 to 52 years, with a mean age of 28 years. Highest level of schooling achieved ranged from grade 7 to grade 12 plus vocational training. None of the subjects in this group repeated any grades or received any extra help during school. All but two of the adults were employed at time of testing (three laborers, four waiter/waitresses, two certified nursing assistants). All were monolingual speakers of English. None of the subjects appeared to have any sensory or intellectual deficiencies. None of the subjects reported a history of speech/ language problems and no subject received any speech/language services. The mean score on the Producing Formulated Sentences subtest of the Clinical Evaluation of Language Functions test was 94 percent (ranging from 83 percent to 100 percent). Most of the errors made by these adults consisted of incomplete sentences (for example, "Only if you are in my s h o e s . . . "). Materials
Forty-eight object names were selected based on the following characteristics: 1) frequency of name (high vs. low); and 2) length of name (short vs. long). Frequency of the name of each object was determined using Carroll, Davies, and Richman's (1971) word frequency count. The name of an object was classified as "High" if it occurred at a frequency rate of at least one per million and "Low" if it occurred at a frequency rate of less than one per million. Length was defined using the number of syllables in the name of an object. Objects with one- or two-syllable names were said to be "Short" names whereas three- or four-syllable names were said to be "Long" names. Using these two characteristics, four groups of object names were created (with mean frequency of occurrence given here): 1) High frequency/Short names (5.275); 2) High frequency/Long names (5.267); 3) Low frequency/ Short names (0.419); and, 4) Low frequency/Long names (0.440). Each group contained twelve object names. No significant differences ex-
186
OmGi1vs, PAvrE~avS,ANY PRoa~voses
isted in the frequency of occurrence between "Short" and "Long" object names for either "Low" frequency, F(1,23) = 0.045, or "High" frequency, F(1,23) = 0.000, object names. All 48 stimuli were used in each of the four tasks in the study (see Appendix). For three of the tasks (object recognition, object naming, and spelling) a photograph was taken of each object. For the objectrecognition task, photographs of the twelve objects in each group were divided into three sets of four photographs. This procedure resulted in a total of twelve sets of photographs. Photographs from each set were mounted on paper in a 2 × 2 arrangement. For the fourth task, the reading task, the name of each object was printed on a separate 3" × 5" index card.
Procedure Each adult was tested individually in one session which ranged from 30 to 60 minutes. Each adult performed four tasks in the following order: 1) spelling task; 2) object-naming task; 3) reading task; and, 4) object-recognition task. For each task the stimuli were presented in a random order, but the presentation order for each task remained fixed across subjects. Spelling Task. Photographs of objects were presented one at a time. The subject was asked to print the name of the object in the photograph on a sheet of paper containing blanks numbered 1 to 48. Object-Naming Task. The procedure for this task was similar to the spelling task except the subject was asked to name the object in the photograph. When an incorrect response was given, the subject was asked to provide another name for the object. The subject was allowed three trials to name an object correctly. The experimenter recorded all responses. Reading Task. In this task the subject was presented with index cards one at a time and asked to read the word printed on the card. The experimenter recorded all responses. Object-Recognition Task. The subject was shown a set of four photographs. The experimenter named one of the objects. The subject was asked to point to the object named by the experimenter.
Results Object Naming To determine if adults with written-language problems also experience object-naming difficulties, performance on the object-recognition and object-naming tasks was compared for each group. For each subject, any object not recognized was eliminated from the object-
OBJECT-NAMINGASXLXTYOr ADULTS
187
naming calculations to ensure that any differences between the groups could not be attributed to a vocabulary deficit. For the experimental group a mean number of two objects per subject was eliminated. For the control group, no objects were eliminated from the analysis because all subjects recognized all the objects. Scores were converted to percentage correct. In the object-naming task subjects were allowed three trials to name an object accurately. The mean number of objects recognized and objects n a m e d correctly on the first trial by each group is presented in Table I. A 2 x 2 split-plot ANOVA was conducted, involving Group (experimental, control) by Task (object recognition, object naming) with repeated measures on the last factor. The analysis yielded a significant main effect of Group, F(1,20) = 18.698, p = .0006, and of Task, F(1,20) = 145.637, p
