Handbook of Clinical Neurology, Vol. 111 (3rd series) Pediatric Neurology Part I O. Dulac, M. Lassonde, and H.B. Sarnat, Editors © 2013 Elsevier B.V. All rights reserved
Specific language impairment ALAN G. KAMHI* AND MARY KRISTEN CLARK Department of Communication Sciences and Disorders, University of North Carolina, Greensboro, NC, USA
The acquisition of language is one of the most important achievements in young children, in part because most children appear to acquire language with little effort. Some children are not so fortunate, however. Children with autism, intellectual disabilities, hearing loss, neurological damage, and those who suffer severe environmental deprivation have obvious reasons for their language learning difficulties. There is a large group of children who also have difficulty learning language, but do not have obvious neurological, cognitive, sensory, emotional, or environmental deficits. Children with language disorders have been variously referred to as language disordered, language impaired, language delayed, or as having a specific language impairment (SLI). Clinicians tend to use the first three terms; SLI is the preferred term in research publications. A language disorder can be defined as a significant delay in the use and/or understanding of spoken or written language. The disorder may involve the form of language (phonology, syntax, and morphology), its content or meaning (semantics), or its use (pragmatics), in any combination (American Speech–Language– Hearing Association, 1993). Children with SLI have intrigued researchers for many years because there is no obvious reason for their language learning difficulties (Leonard, 1998). SLI has been found to be an enduring condition that begins in early childhood and often persists into adolescence and adulthood (Stothard et al., 1998; Tomblin et al., 2003). The language problems of children with SLI are not limited to spoken language; they also affect reading and writing and thus much of academic learning (Catts et al., 2005). At least half of kindergarten children with language disorders have identifiable reading and learning difficulties in later primary grades (Catts et al., 2002) and continue to demonstrate decreased reading achievement compared to typical language peers
through at least Grade 10 (Catts et al., 2008). Language abilities are also basic to everyday functioning, so it is not surprising to find that one-third of adolescents with SLI are less independent than typically developing peers (Conti-Ramsden and Durkin, 2008).
CLASSIFICATION AND CHARACTERISTICS Children with language disorders can be classified according to the aspect of language that is impaired (phonology, syntax, morphology, semantics, and/or pragmatics), its severity (mild, moderate, or severe), and whether it affects comprehension (receptive language), production (expressive language), or both (Bishop, 1997). The expressive–receptive distinction is used by the Diagnostic and Statistical Manual of Mental Disorders-IV (DSM-IV; American Psychiatric Association, 1994) and the International Classification of Functioning, Disability, and Health (updated 2011). According to the DSM-IV, an expressive language disorder is characterized by language production skills that are below an age-appropriate level. Deficits in expressive language are apparent in both the formation and production of language and include weak vocabulary skills, word finding difficulties, word omissions, poor narrative skills (i.e., storytelling), difficulty providing information, and grammatical errors. Expressive language disorder can affect both oral and written language. Children with expressive language disorder demonstrate age-appropriate comprehension of language. Deficits in receptive language are apparent when children do not understand language at an age-appropriate level and include difficulties following directions, understanding words and sentences, and answering questions. Comprehension of both oral and written language can be affected. Mixed receptive–expressive
*Correspondence to: Alan G. Kamhi, Department of Communication Sciences and Disorders, 300 Ferguson, UNCG, Greensboro, NC 27402, USA. E-mail: [email protected]
A.G. KAMHI AND M.K. CLARK
language disorder is characterized by difficulties with both the production and comprehension of language. Subgroups of children with language disorders have been identified according to the expressive–receptive distinction. In a group of kindergarten children, 35% had expressive problems, 28% had receptive problems, and 35% both expressive and receptive problems (Tomblin et al., 1996). Others have identified additional subtypes of language impairment and shown that almost half of children moved from one subtype to another over only a year as their specific strengths and weaknesses changed (Conti-Ramsden et al., 1997; Conti-Ramsden and Botting, 1999). Leonard (2009) has recently raised serious questions about whether expressive language disorder is an accurate diagnostic category. In a provocative article, he argues that close inspection of the evidence indicates that limitations in language knowledge or processing language input typically accompany deficits in language expression. There is no theory of expressive language problems that does not assume a limitation in language knowledge or a problem in processing language input. In addition, it has proven difficult to test the specific correlates of an expressive language problem in comprehension. In short, despite the common practice of categorizing children with language disorders as having expressive and/or receptive language problems, there appears to be little evidence that a pure expressive language disorder exists. Another way to classify children with language disorders is to differentiate children with SLI from children with non-specific language impairments (NLI). Children with NLI do not perform within normal age limits on measures of nonverbal intelligence. Research comparing SLI and NLI has found that children with NLI are just as responsive to therapy as children with SLI (Fey et al., 1994). More recent research has found that the risk for later reading disabilities is higher for children with NLI (nonverbal and language deficits) than for those with SLI (language deficits only; Catts et al., 2002).
IDENTIFYING CHILDREN WITH SLI Children with SLI are identified using a combination of exclusionary and inclusionary criteria. In addition to excluding children with intellectual disabilities, hearing loss, severe emotional disturbance, and frank neurological deficits, children with severe phonological impairments are also excluded, but the specific criterion varies. In some cases, children whose speech delays are more severe (6 months or more) than their language difficulties are excluded (Aram et al., 1993) whereas in others, children are only excluded if their speech errors affect performance in the study. For example, in studies
involving grammatical morphology, children must be able to produce final /s/ and /t/ (Goffman and Leonard, 2000; Leonard et al., 2006). Although many definitions of SLI include environmental factors as an exclusionary criterion, most studies do not exclude children from low socioeconomic backgrounds. The principal inclusionary criterion is performance within the normal range on a measure of nonverbal intelligence. Most investigators define “typical performance” as within one standard deviation (SD) of the mean (i.e., nonverbal IQ must be above 85). Some investigators allow nonverbal IQ to be as low as 70 to allow for measurement error whereas others (e.g., Plante, 1998) use 75 as the cut-off point to clearly differentiate children with SLI from those with significant intellectual disabilities (IQ < 68–70). Tomblin et al. (1996) attempted to standardize the criteria used for diagnosing SLI in a prospective study of more than 7000 kindergarten children in the state of Iowa (US). The diagnostic sensitivity for SLI was found to be best when two or more composite scores of language modality (comprehension and production) or domain (vocabulary, grammar, narration) were more than 1.2 standard deviations below the mean. Children also had to meet the usual exclusionary criteria and perform within normal age limits on a measure of nonverbal intelligence. Using these criteria, 7.4% of the kindergarten children met the criteria for SLI. The exclusionary and inclusionary criteria used by researchers to define SLI are usually not used by clinicians to identify children with language impairments. Clinical assessments are driven more by service delivery concerns than classification or etiology issues. Standardized measures of expressive and receptive language are typically used by clinicians to identify children who need services. In many clinical settings, a child who performs more than 1 SD below age level on a norm-referenced measure of language is typically eligible for services. Clinicians do not routinely administer measures of nonverbal intelligence because performance on these tests does not influence eligibility for services. Assessing intellectual abilities is also viewed as being outside of the scope of practice of speech–language pathologists, even though some tests were developed to be used by clinicians as well as psychologists (e.g., The Test of Nonverbal Intelligence — Third Edition; Brown et al., 1997).
LANGUAGE CHARACTERISTICS OF CHILDREN WITH SLI The language abilities of children with SLI are commonly viewed as being delayed relative to typical developing children. A straightforward delay would mean that a 5-year-old child with SLI would exhibit language
SPECIFIC LANGUAGE IMPAIRMENT abilities similar to a 3-year-old typically developing child. With successful intervention, this 5-year-old child would catch up to his typically developing peers within a couple of years and show no further consequences of having a language disorder. A delay thus suggests a late start and nothing more. SLI is not a language delay. The language abilities of a 5-year-old child with SLI do not mirror the language abilities of a younger typically developing child. Their syntactic and morphological abilities may be comparable to younger children, but their semantic and pragmatic abilities are more in line with age-level peers (Leonard, 1998). The 2-year age difference affects what the 5-year-old child talks about (semantics) and their ability to use language (pragmatics). Moreover, many children with SLI are at risk for subsequent academic and learning problems (Catts et al., 2005), so even if their spoken language abilities normalize, languagebased learning problems persist. If the casual observer were to listen to a 5-year-old child with SLI, they would notice that sentences are generally shorter and less complex than their age peers. They would also notice some difficulty understanding complex sentences, questions, and following longer narrative discourses. Grammatical morpheme errors, such as incorrect past tense forms, auxiliary/copula omissions, and pronoun errors would be very noticeable (Leonard, 1998). Pragmatic and semantic (lexical) differences may not be noticeable to the casual observer, though difficulties in these areas may exist in some children given the heterogeneity of the disorder.
CAUSES AND CONSEQUENCES There has been considerable interest in explaining why children with normal nonverbal intelligence, intact speech and hearing mechanisms, and supportive language learning environments have difficulty acquiring language. In the sections below, we consider the genetic and neurobiological aspects of SLI and then review various explanatory accounts of the disorder.
Genetic factors All developmental disorders have a strong genetic component, and SLI is no exception. Studies have shown that between 20% and 40% of children with language impairments have an affected family member (Stromswold, 1998). Twin studies confirm that these numbers reflect genetic factors rather than the shared environment of the family members (Bishop, 2002). The idea of a gene for language received considerable attention in the 1990s with the discovery of the KE family in London. Sixteen members of this three-generational family were diagnosed with a language disorder (Gopnik and Crago, 1991). All of the affected individuals had a mutation
of a gene on chromosome 7 referred to as FOXP2. The initial excitement of the discovery of the language gene was muted by subsequent research showing that the gene regulated the activity of other genes and affected the development of many organs including brain systems involved in speech and language learning (Fisher, 2005). It also turned out that most individuals with SLI have an intact FOXP2 gene. This finding does not reduce the impact genetics has on language disorders; it just means that language disorders, like other developmental disabilities, are caused by several genes as well as environmental factors.
Neurobiological factors Neuroimaging studies have attempted to identify morphometric factors that differentiate individuals with SLI from those with normal language, in particular whether children with SLI have atypical asymmetry of the perisylvian region which includes both Broca’s and Wernicke’s areas. Findings from these studies have been inconsistent. Some MRI studies have found reduced or reversed volumetric asymmetry of a number of perisylvian structures, including the planum temporale, the pars triangularis, and the inferior frontal gyrus (Cohen et al., 1989; Plante et al., 1991; Ja¨ncke et al., 2007). Functional imaging studies have also found differences in activation. Children with SLI showed a greater propensity for symmetrical cortical activation while performing language tasks (Bernal and Altman, 2003) and during the resting state (Ors et al., 2005). However, other studies have failed to replicate these findings (Preis et al., 1998) or have found considerable variability between children (Shafer et al., 2000). A recent study by Whitehouse and Bishop (2008) attempted to address some of the methodological limitations in previous studies. Using Doppler ultrasonography, they compared activation and cerebral dominance for language function in young adults with SLI with three comparison groups: (a) adults with a history of childhood SLI, (b) adults with autism spectrum disorder (ASD) and comorbid language impairment, and (c) adults with no history of a developmental disorder. The findings were quite remarkable. All of the participants in the SLI-history group and the majority of participants in the ASD (82%) and typical (91%) groups had greater activation in the left cerebral arteries. In contrast, the majority of young adults in the SLI group had language function lateralized to the right hemisphere (55%) or showed bilateral functioning (27%). Given the strong genetic etiology of SLI, Whitehouse and Bishop (2008) suggest that genes that increase the risk for SLI may also disrupt the neurodevelopmental processes that establish cerebral lateralization.
A.G. KAMHI AND M.K. CLARK
The argument “is not that atypical cerebral lateralization is pathological in itself, but rather that there are pathological processes that lead to atypical cerebral lateralization, and at the same time impair language development” (p. 3198). Atypical cerebral dominance does not occur in all cases of poor language development (i.e., ASD and SLI-history groups), but may act as a biological marker of persisting SLI.
EXPLANATORYACCOUNTS OF SLI In addition to research examining genetic and neurobiological factors in SLI, researchers have spent the last three decades examining the linguistic and cognitive deficiencies that may account for the specific language learning problems experienced by individuals with SLI. If the deficit is truly specific to language, this would suggest either a language-specific processing problem or a deficit limited to grammatical rule learning. If the learning problem is not specific or limited to language, then language learning problems may be explained by some type of general processing limitation. In the sections to follow, the most prominent linguistic and processing accounts of SLI are considered. The section concludes with a recent attempt to provide a unified explanation for SLI.
LINGUISTIC ACCOUNTS Proponents of language knowledge deficit accounts believe that grammatical deficits are caused by incomplete knowledge of particular rules, principles, or constraints (Leonard, 1998). More specifically, it has been proposed that children with SLI have difficulty establishing structural agreement relations (Clahsen, 1989), have problems acquiring implicit grammatical rules (Gopnik and Crago, 1991), go through a protracted period in which they fail to consistently mark tense in main clauses (Rice et al., 1995), or have a representational deficit for dependent relations (van der Lely and Christian, 2000). Clahsen’s (1989) missing agreement account was one of the first attempts to specify the nature of the linguistic deficit in children with SLI. According to this account, grammatical deficits are due to a selective impairment in establishing structural relations of agreement. Aspects of language that would be affected by an agreement deficit include verb inflections, auxiliaries and copulas, gender and number, and case markers on determiners. The difficulty many children with SLI have with these forms is consistent with this account, but some data cannot be explained by agreement problems. Roberts (1995; as cited in Leonard, 1998), for example, found that the German-speaking children with SLI in Clahsen’s study showed higher percentages of appropriate verb agreement inflections and copula forms than
English-speaking children with SLI. The missing agreement account cannot explain why the German children had fewer agreement errors than English-speaking children given that these forms must be memorized. Gopnik and her colleagues (e.g., Gopnik, 1990; Gopnik and Crago, 1991) believe that children with SLI have difficulty acquiring implicit grammatical rules involving tense, number, and person. This difficulty forces children either to memorize inflected forms as unanalyzed lexical items or learn explicit rules such as “add –s for more than one” (Gopnik, 1994). There is evidence, however, that children with SLI can learn implicit rules (Swisher et al., 1995). This account also does not explain why the rule-learning deficit affects children learning Hebrew and Italian less than it affects children learning English. The extended optional infinitive (EOI) hypothesis is based on the work of Wexler (1994). Wexler provided evidence that young children go through a stage during which they fail to obligatorily mark tense in main clauses, even though they know the grammatical properties of finiteness (i.e., marking tense). Children with SLI stay in this extended optional infinitive stage for a much longer time than typically developing children. Indeed, there is some evidence that some children with SLI never leave this state (Rice et al., 1995). Although these children know the rules for adding grammatical morphemes to the ends of words, they do not know that their use is obligatory. In their grammars, both nonfinite (infinitive) and finite (marked for tense and agreement) forms are correct. There is considerable evidence that is consistent with the EOI view (Leonard, 1998). This account also provides an explanation for the frequent occurrence of pronominal case errors and problems with verb morphology (Leonard, 1995). Like other linguistic accounts, the EOI hypothesis has the most difficulty with some cross-linguistic data. For example, Hebrew- and Italian-speaking children with SLI produce finite forms at the same rate as language-matched children. The EOI hypothesis also cannot explain the high frequency of substitution errors (e.g., was for were) and some past tense errors (Leonard et al., 2007). Leonard et al. have shown that verbs with a durative aspect (e.g., play) are less likely to be marked with past tense than verbs with nondurative aspect (e.g., drop). The final linguistic account to be considered is van der Lely’s representational deficit view (e.g., van der Lely, 2005; van der Lely and Christian, 2000). The account is different from previous linguistic explanations because it attempts to explain why children with SLI show poor grammatical comprehension as well as poor grammatical production abilities. It is similar to other accounts in its attempt to explain the difficulty children
SPECIFIC LANGUAGE IMPAIRMENT with SLI have with agreement, tense, and case. According to van der Lely, these problems are caused by a representational deficit for dependent relationships. This view does not appear to be as well specified as previous accounts and has problems explaining some findings in English as well as non-Germanic languages (Leonard, 1998). Evaluation of linguistic accounts Language knowledge deficit accounts have advanced the precision of research in the study of SLI and provided plausible explanations for the grammatical difficulties experienced by children with SLI (Leonard, 1998; Ullman and Pierpont, 2005). The different linguistic frameworks have shown how seemingly disparate grammatical forms (e.g., nominative case and auxiliary be forms) may be connected to a particular linguistic deficit. As Leonard (1998, p. 235) notes, however, the Achilles’ heel of these accounts is that their predictions hold for a limited range of languages. And even within grammar, these accounts cannot fully explain the combination of syntactic, morphological, and phonological deficits that occur (Ullman and Pierpont, 2005). More important is that linguistic explanations cannot account for the nonlinguistic difficulties experienced by children with SLI. The next section considers these deficits.
GENERAL PROCESSING ACCOUNTS The term SLI and the numerous language deficit accounts to explain the disorder are consistent with the assumption that SLI is an impairment restricted to language. There is a substantial literature, however, showing that children with SLI experience difficulty in a variety of nonlinguistic abilities involving mental representation, mental imagery, hypothesis testing, analogical reasoning, and hierarchical planning (for reviews, see Bishop, 1992 or Leonard, 1998). The fact that children with SLI perform within normal limits on nonverbal tests of intelligence makes the existence of these nonlinguistic deficits somewhat perplexing. Johnston (1982) provides a solution to the puzzle. Nonverbal intelligence tests contain items that require visual perception of static figures, shapes, and designs. Children with SLI obviously do well in tasks that have limited representational and memory demands, but their poorer performance on more demanding tasks suggests that they might have a more general processing deficit that impacts performance on certain nonlinguistic tasks as well as language. The two most likely candidates for a general processing deficit are working memory and speed of processing. Limitations in processing speed or working memory would adversely affect a child’s ability to generate grammatical rules from language input as well as affect the use of
these rules once they are acquired. For example, children with SLI may have no difficulty generating a rule that a grammatical inflection such as —ed refers to past tense, but have difficulty processing the continuous stream of speech quickly enough to identify the morpheme, determine its grammatical function, and store it before having to devote attention to the portion of the speech stream that follows (Leonard et al., 2007). The initial proposal that SLI could be explained as a deficit in general speed of processing came from the work of Kail (1994) who found that reaction times (RTs) of children with SLI were uniformly slower than those of typically developing children regardless of the task or domain. Analyzing data across several studies, Kail found that children with SLI were 33% slower across a range of language and nonlanguage tasks. Subsequent studies found the generalized slowing to be lower than Kail’s initial finding; 18% in Windsor and Hwang (1999) and 14% in Miller et al. (2001). Verbal short-term and working memory deficits are well documented in children with SLI (Archibald and Gathercole, 2006; Montgomery and Evans, 2009). Many studies have used the working memory model of Baddeley (1986) which includes a central executive and two modality-specific storage systems: a phonological loop and a visuospatial sketchpad. Gathercole and Baddeley (1990) proposed that children with SLI have particular difficulty with phonological memory because they had difficulty repeating nonwords but performed comparably to age peers on discriminating word pairs and articulation rate. Although subsequent studies have confirmed the difficulty children with SLI have in repeating nonwords (Dollaghan and Campbell, 1998; Ellis Weismer et al., 2000), there is also evidence that working memory limitations affect visual processing as well (Hoffman and Gillam, 2004). As Leonard et al. (2007) note, evidence supporting a broad-based processing deficit would have significant implications because it would suggest that SLI is not strictly a language disorder. To investigate this possibility, Leonard et al. (2007) examined the influence of processing speed and working memory in 204 14-year-old children with SLI. Speed of processing and working memory limitations accounted for 62% of the variance in children’s composite language scores. Verbal working memory was more important than speed of processing, but speed did contribute significantly to language performance. Confirmatory factor analysis models indicated that speed and working memory were not interchangeable measures of processing; they were functionally separable. It was also possible to further divide speed into motor speed, nonlinguistic cognitive speed, and linguistic speed. Similarly, working memory could be further divided into
A.G. KAMHI AND M.K. CLARK
verbal and nonverbal working memory. The most important finding that emerged from these subdivisions was that nonlinguistic/nonverbal abilities functioned as separate dimensions. A comprehensive account of language impairment thus needs to explain deficits in nonlinguistic motor and cognitive factors as well as language-based deficits. Evaluation of general processing accounts In a discussion of their findings, Leonard et al. (2007) consider some alternative ways to view the relationship between processing measures and language. The most extreme view is that a processing limitation severely reduces performance on language tests and possibly masks the knowledge children actually possess. The dramatic improvement children with SLI made following a month of treatment designed to improve auditory temporal processing abilities (Tallal et al., 1996) supports this view. A more likely explanation, however, is that poor language scores are affected by processing limitations as well as language knowledge deficiencies. Unfortunately, it has proven difficult to rule out the influence of processing limitations in the acquisition of language knowledge. Tomasello (2003), for example, has argued that working memory limitations in young children could explain the extended optional finite verb use. Limited language knowledge could also be the result of a lifetime of functioning with limited language processing skills. The burden of proof appears to lie with those who advocate non-processing explanations for the language deficit.
SPECIFIC PROCESSING ACCOUNTS Whereas general processing accounts attribute deficits to domain-general processes, specific processing accounts postulate domain-specific processing limitations. The most prominent explanation has focused on auditory temporal processing abilities. In brief, the work of Tallal and her colleagues over the past 30 years has shown that compared to age-matched peers, SLI children have difficulty perceiving and producing information rapidly in time. Renamed a “neural timing deficit” in Tallal (1988), this processing deficit accurately predicts (a) performance on nonverbal and verbal serial memory tasks, (b) the pattern and type of speech production and perception errors, (c) the degree of receptive language impairment, and (d) classification of SLI. In 1996, Tallal and Mersenich (Merzenich et al., 1996; Tallal et al., 1996) developed a computerized intervention program using acoustically modified speech (Fast ForWord) to remediate the proposed processing deficit. Although their initial studies showed remarkable advances in response to the intensive treatment, recent studies have shown that
children made similar gains with other intensive treatments (Gillam et al., 2008). The causal connection between a temporal processing deficit and language learning problems has always been unclear. For example, typically developing (TD) children younger than 4;6 cannot perform the temporal processing tasks used in Tallal’s studies (Tallal, 1976). If a temporal processing disorder causes language learning difficulties, how would one explain the sophisticated language abilities of 4;6-year-old TD children? By this age, TD children produce many complex sentence structures and use a wide range of grammatical morphemes (Owens, 2007). Tallal et al. (1996) recognized this dilemma, noting that it seems unlikely that children with SLI learned the equivalent of 2 years of language after a 1-month intervention. They presumably had already developed considerably more language competence than they were able to demonstrate under normal speaking and listening conditions (p. 83). As Leonard (1998) points out, “this leads to the necessary conclusion that the processing deficit did not hinder language development to the extent generally assumed” (p. 275). Evaluation of specific processing accounts In light of these concerns, the most optimistic interpretation of the data is that temporal processing limitations may account for protracted acquisition of selected language forms, such as grammatical morphemes and structures that depend on grammatical morphemes for their meaning. However, a temporal processing deficit does not explain the problem some SLI children continue to have with these forms throughout childhood and adolescence or the problems with other language forms (e.g., temporal connectives, adverbials, etc.) some of these children experience. The most pessimistic interpretation of the data is that the relationship between these temporal processing abilities and language is not unidirectional (i.e., from perception to language), but interactive and reciprocal. That is, a temporal processing deficit may cause certain language problems, but the language problems may, in turn, cause the temporal processing deficit (Rees, 1981).
A UNIFYING ACCOUNT: THE PROCEDURAL DEFICIT HYPOTHESIS
The most recent attempt to explain SLI is the procedural deficit hypothesis (PDH) proposed by Ullman and Pierpont (2005). As evident from the previous sections, the two broad competing accounts for SLI fall short of explaining the full range of linguistic and nonlinguistic deficits in children with SLI. The major advantage of the PDH is that it provides one unifying explanation for the strengths and weaknesses of children with SLI.
SPECIFIC LANGUAGE IMPAIRMENT The hypothesis is that many if not most children with SLI have abnormalities of brain structure and function involved in the procedural memory system. The affected brain structures are the left hemispheric frontal/basal ganglia circuits and Broca’s area. Ullman and Pierpont cite recent evidence showing the involvement of the procedural memory system in rule-governed aspects of grammar. The declarative memory system, in contrast, is involved in learning the lexicon. The PDH thus explains the relative strengths children with SLI have in lexical learning and their difficulties with rule-based aspects of grammar like morphosyntax. The hypothesis also explains the cognitive strengths (nonverbal intelligence) and weaknesses (mental imagery, working memory, and temporal processing) of children with SLI. The PDH has similarities and differences from the two major competing explanations for SLI. It is similar to language knowledge accounts by acknowledging that grammar is directly affected, but diverges from these accounts by extending the impairment beyond grammar to lexical retrieval and selected nonlinguistic abilities. It is similar to processing deficit accounts in acknowledging that children with SLI often have deficits in working memory and temporal processing. Unlike processing explanations, however, the PDH posits that these deficits reflect an underlying dysfunction in procedural memory that also causes other impairments in grammar, lexical retrieval, and motor functions. Importantly, the PDH differs from previous explanations in three ways: (1) it is highly predictive and testable because it leads to novel predictions about SLI on the basis of independent sources of information; (2) it attempts to explain within a single theoretical framework a large amount of SLI data within and across subjects and thus has the potential to explain the heterogeneity of the disorder; (3) the hypothesis is a hypothesis about brain and behavior, rather than a theory about distinct linguistic or cognitive abilities (Ullman and Pierpont, 2005). Evaluation of the procedural deficit hypothesis Because the PDH is relatively new, it is only beginning to receive attention in the literature. A recent study by Tomblin et al. (2007) examined serial reaction time, as a direct test of adolescents’ procedural learning abilities, in adolescents with and without SLI. The findings were consistent with the PDH. Adolescents with SLI demonstrated significantly slower reaction times and slower learning rates than typically developing children. Reaction times were also significantly correlated to a measure of grammar, but not vocabulary, which is consistent with the PDH. We look forward to future research on this promising explanatory account of SLI.
CLINICAL AND EDUCATIONAL IMPLICATIONS The language and academic learning problems of many children with SLI persist throughout childhood into adolescence and beyond. Identifying these children as early as possible is thus particularly important. Knowledge of the characteristics of SLI should aid physicians, pediatricians, and early childhood specialists to identify these children during the preschool years and ensure that they receive appropriate services. With high quality language intervention and literacy instruction, most children with SLI should be able to perform and function adequately in school and beyond.
REFERENCES American Psychiatric Association (1994). Diagnostic and Statistical Manual of Mental Disorders. 4th edn. American Psychiatric Press, Washington, DC. American Speech–Language–Hearing Association (1993). Definitions of communication disorders and variations. Cited Sept 23 2009. ASHA, Maryland. Available at: http://www.asha.org/docs/pdf/RP1993-00208.pdf. Aram DM, Morris R, Hall NE (1993). Clinical and research congruence in identifying children with specific language impairment. J Speech Hear Res 36: 580–591. Archibald LMD, Gathercole SE (2006). Short term and working memory in specific language impairment. Int J Lang Commun Disord 41: 675–693. Baddeley A (1986). Working Memory. Oxford University Press, Oxford. Bernal B, Altman AR (2003). Speech delay in children: a functional MR imaging study. Radiology 229: 651–658. Bishop DVM (1992). The underlying nature of specific language impairment. J Child Psychol Psychiatry 33: 3–66. Bishop DVM (1997). Uncommon Understanding: Development and Disorders of Language Comprehension in Children. Psychology Press, East Sussex. Bishop DVM (2002). The role of genes in the etiology of specific language impairment. J Commun Disord 35: 311–328. Brown L, Sherbenou RJ, Johnsen SK (1997). Test of Nonverbal Intelligence. 3rd edn. Pro-Ed, Texas. Catts HW, Fey ME, Zhang X et al. (2002). A longitudinal investigation of reading outcomes in children with language impairments. J Speech Lang Hear Res 45: 1142–1157. Catts HW, Adlof SM, Hogan TP et al. (2005). Are specific language impairment and dyslexia distinct disorders? J Speech Lang Hear Res 48: 1378–1396. Catts HW, Bridges MS, Little TD et al. (2008). Reading achievement growth in children with language impairments. J Speech Lang Hear Res 51: 1569–1579. Clahsen H (1989). The grammatical characterization of developmental dysphasia. Linguistics 27: 897–920. Cohen M, Campbell R, Yaghmai F (1989). Neuropathological abnormalities in developmental dysphasia. Ann Neurol 25: 567–570.
A.G. KAMHI AND M.K. CLARK
Conti-Ramsden G, Botting N (1999). Classification of children with specific language impairment: longitudinal considerations. J Speech Lang Hear Res 42: 1195–1204. Conti-Ramsden G, Durkin K (2008). Language and independence in adolescents with and without a history of specific language impairment (SLI). J Speech Lang Hear Res 51: 70–83. Conti-Ramsden G, Crutchley A, Botting N (1997). The extent to which psychometric tests differentiate subgroups of children with SLI. J Speech Lang Hear Res 40: 765–777. Dollaghan C, Campbell T (1998). Nonword repetition and child language impairment. J Speech Lang Hear Res 41: 1136–1146. Ellis Weismer S, Tomblin JB, Zhang X et al. (2000). Nonword repetition performance in school-age children with and without language impairment. J Speech Lang Hear Res 43: 865–878. Fey ME, Long SH, Cleave PL (1994). Reconsideration of IQ criteria in the definition of specific language impairment. In: RV Watkins, ML Rice (Eds.), Specific Language Impairments in Children. Paul H. Brookes, Maryland, pp. 161–178. Fisher SE (2005). Dissection of molecular mechanisms underlying speech and language disorders. Appl Psycholinguist 26: 111–128. Gathercole S, Baddeley A (1990). Phonological memory deficits in language impaired children: is there a causal connection? J Mem Lang 29: 336–360. Gillam RB, Loeb DF, Hoffman LM et al. (2008). The efficacy of Fast ForWord language intervention in school-age children with language impairment: a randomized controlled trial. J Speech Lang Hear Res 51: 97–119. Goffman L, Leonard J (2000). Growth of language skills in preschool children with specific language impairment: implications for assessment and intervention. Am J Speech Lang Pathol 9: 151–161. Gopnik M (1990). Feature-blind grammar and dysphasia. Nature 344: 715. Gopnik M (1994). The family. McGill Working Papers in Linguistics 10: 1–4. Gopnik M, Crago MB (1991). Familial aggregation of a familial language disorder. Cognition 39: 1–50. Hoffman L, Gillam RB (2004). Verbal and spatial information processing constraints in children with specific language impairment. J Speech Lang Hear Res 47: 114–125. International Classification of Functioning, Disability, and Health. (2011). Cited Sep 23 2009. World Health Organization, Geneva. Available at: http://apps.who.int/ classifications/icfbrowser/. Ja¨ncke L, Siegenthaler T, Preis S et al. (2007). Decreased white-matter density in a left-sided fronto-temporal network in children with developmental language disorder: evidence for anatomical anomalies in a motor-language network. Brain Lang 102: 91–98. Johnston JR (1982). Interpreting the Leiter IQ: performance profiles of young normal and language-disordered children. J Speech Hear Res 25: 291–296. Kail R (1994). A method of studying the generalized slowing hypothesis in children with specific language impairment. J Speech Hear Res 37: 418–421.
Leonard LB (1995). Functional categories in the grammars of children with specific language impairment. J Speech Hear Res 38: 1270–1283. Leonard LB (1998). Children with Specific Language Impairment. MIT Press, Massachusetts. Leonard LB (2009). Is expressive language disorder an accurate diagnostic category? Am J Speech Lang Pathol 18: 115–123. Leonard LB, Camarata SM, Pawlowska M et al. (2006). Tense and agreement morphemes in the speech of children with specific language impairment during intervention: phase 2. J Speech Lang Hear Res 49: 749–770. Leonard LB, Ellis Weismer S, Miller CA et al. (2007). Speed of processing, working memory, and language impairment in children. J Speech Lang Hear Res 50: 408–428. Merzenich M, Jenkins W, Johnston P et al. (1996). Temporal processing deficits of language-learning impaired children ameliorated by training. Science 271: 77–81. Miller C, Kail R, Leonard LB et al. (2001). Speed of processing in children with specific language impairment. J Speech Lang Hear Res 44: 416–433. Montgomery JW, Evans JL (2009). Complex sentence comprehension and working memory in children with specific language impairment. J Speech Lang Hear Res 52: 269–288. Ors M, Ryding E, Lindgren M et al. (2005). SPECT findings in children with specific language impairment. Cortex 41: 316–326. Owens R (2007). Language Development. 6th edn. Allyn & Bacon, Massachusetts. Plante E (1998). Criteria for SLI: the Stark and Tallal legacy and beyond. J Speech Lang Hear Res 41: 951–957. Plante E, Swisher L, Vance R et al. (1991). MRI findings in boys with specific language impairment. Brain Lang 41: 52–66. Preis S, Ja¨ncke L, Schittler P et al. (1998). Normal intrasylvian anatomical asymmetry in children with developmental language disorder. Neuropsychologia 36: 849–855. Rees N (1981). Saying more than we know: is auditory processing disorder a meaningful concept? In: R Keith (Ed.), Central Auditory and Language Disorders in Children. College-Hill Press, California, pp. 94–120. Rice M, Wexler K, Cleave P (1995). Specific language impairment as a period of extended optional infinitive. J Speech Hear Res 38: 850–863. Shafer VL, Schwartz RG, Morr ML et al. (2000). Deviant neurophysiological asymmetry in children with language impairment. Neuroreport 11: 3715–3718. Stothard SE, Snowling MJ, Bishop DVM et al. (1998). Language impaired preschoolers: a follow-up into adolescence. J Speech Lang Hear Res 41: 407–418. Stromswold K (1998). Genetics of spoken language disorders. Hum Biol 70: 297–324. Swisher L, Restrepo MA, Plante E et al. (1995). Effect of implicit and explicit “rule” presentation on bound-morpheme generalization in specific language impairment. J Speech Hear Res 38: 168–173. Tallal P (1976). Rapid auditory processing in normal and disordered language development. J Speech Hear Res 19: 561–571.
SPECIFIC LANGUAGE IMPAIRMENT Tallal P (1988). Developmental language disorders. In: J Kavanagh, T Truss (Eds.), Learning Disabilities: Proceedings of the National Conference. York Press, Maryland, pp. 181–272. Tallal P, Miller S, Bedi G et al. (1996). Language comprehension in language-learning impaired children improved with acoustically modified speech. Science 271: 81–84. Tomasello M (2003). Constructing a Language: A Usagebased Theory of Language Acquisition. Harvard University Press, Massachusetts. Tomblin JB, Records NL, Zhang X (1996). A system for the diagnosis of specific language impairment in kindergarten children. J Speech Hear Res 39: 1284–1294. Tomblin JB, Zhang X, Buckwalter P et al. (2003). The stability of primary language disorder. J Speech Lang Hear Res 46: 1283–1296. Tomblin JB, Mainela-Arnold E, Zhang X (2007). Procedural learning in adolescents with and without specific language impairment. Lang Learning Dev 3: 269–293.
Ullman MT, Pierpont EI (2005). Specific language impairment is not specific to language: the procedural deficit hypothesis. Cortex 41: 399–433. van der Lely HKJ (2005). Domain-specific cognitive systems: insight from grammatical-SLI. Trends Cogn Sci 9: 53–59. van der Lely HKJ, Christian V (2000). Lexical word formation in children with grammatical SLI: a grammar-specific versus an input-processing deficit? Cognition 75: 33–63. Wexler K (1994). Optional infinitives, head movement, and the economy of derivations. In: D Lightfoot, N Hornstein (Eds.), Verb Movement. Cambridge University Press, New York, pp. 305–350. Whitehouse AJO, Bishop DVM (2008). Cerebral dominance for language function in adults with specific language impairment or autism. Brain 131: 3193–3200. Windsor J, Hwang M (1999). Testing the generalized slowing hypothesis in specific language impairment. J Speech Lang Hear Res 42: 1205–1218.