Specific language impairment (SLI) is a term that is applied to children who show a significant deficit in their spoken language ability with no obvious accompanying problems such as hearing impairment, mental retardation, or neurological damage. This type of language disorder is regarded as developmental in nature because affected children exhibit language learning problems from the outset.

Although SLI is receiving increased attention in the research and clinical literature, it is not a newly discovered disorder. Children meeting the basic definition of SLI have been described in the literature since the 1800s, but have been given a wide range of clinical labels. More recent clinical labels used for children with SLI include developmental aphasia, developmental dysphasia, and developmental language disorder. The last continues to be used in the DSM-IV classification system, with the subtypes of “expressive” and “receptive and expressive” (American Psychiatric Association, 1994). These subtypes acknowledge that some children with SLI may have significant limitations primarily in the area of language production, whereas others may have major limitations in both the comprehension and production of language.

The prevalence of SLI is estimated to be approximately 7% among 5-year-olds, based on epidemiological data (Tomblin et al., 1997). Males outnumber females; the most recent evidence suggests a ratio of approximately 1.5:1.

Children with SLI are two to three times more likely than typically developing children to have parents or siblings with a history of language problems (Tallal, Ross, and Curtiss, 1989; Tomblin, 1989; Tallal et al., 2001). For children with family histories of language problems, there is reason to suspect a genetic basis rather than a primary environmental basis. Concordance rates for SLI are considerably higher for monozygotic twins than for same-sex dizygotic twins (Bishop, North, and Donlan, 1995). Rapid progress is being made in the genetic study of SLI. For a well-studied three-generational family that includes a high proportion of members with SLI, the evidence implicates a region on the long arm of chromosome 7 (Fisher et al., 1998). The same region was identified in a group of children with SLI who participated in an epidemiological study (Tomblin, 1999). However, other recent studies of clinically referred cases of SLI have revealed prominent areas of linkage on chromosomes 16 and 19, but not on chromosome 7 (SLI Consortium, 2002). Clearly, further refinement is needed before the genetic basis for SLI is fully understood (see Bishop, 2002).

In recent years, neuroanatomical evidence of differences between individuals with SLI and typically developing individuals has appeared in the literature (see Ahmed, Lombardino, and Leonard, 2000, for a recent review). The specific differences observed have varied across studies. For example, symmetry of the right and left perisylvian areas seems to be more likely in children with SLI than in controls. Interestingly, this pattern can also be seen in parents or siblings of children with SLI even when they do not exhibit a language disorder. Other studies have revealed a higher likelihood of atypical neuroanatomical patterns in children with SLI than in controls, but differences among the children with SLI in the particular pattern seen (e.g., ventricular enlargement, central volume loss).

Although a diagnosis of SLI is not given to children unless they meet the criteria noted above, many children with SLI nevertheless show subtle weaknesses in other areas. For example, as a group, these children are slower and less accurate on nonlinguistic cognitive tasks such as mental rotation (e.g., Miller et al., 2001), and less coordinated than their typically developing same-age peers (Powell and Bishop, 1992; Hill, 2001). These findings have led to proposals about the possible causes of SLI, but the presence of children with SLI who show none of these accompanying weaknesses raises the possibility that SLI and subtle cognitive and motor weakness are comorbid. That is, the conditions that cause SLI frequently co-occur with conditions that cause these other problems, but the latter are not responsible for SLI.

For children with SLI whose language problems are still present at 5 years of age, difficulties with language may continue into adolescence and even adulthood (Bishop and Adams, 1990; Beitchman et al., 1996). Comparisons of young adults with a history of SLI and same-age adults with no such history reveal differences favoring the latter on a range of spoken production and comprehension tasks (Tomblin, Freese, and Records, 1992).

Children with SLI are at greater risk for reading deficits than children with typical language development. This observation can be explained in part by the fact that children with SLI and those with developmental dyslexia are overlapping populations (McArthur et al., 2000). For example, prospective study of children from homes with a positive history of dyslexia reveals significantly more difficulties with spoken language than children with no such family history (Scarborough, 1990).

The language difficulties experienced by children with SLI cover most or all areas of language, including vocabulary, morphosyntax, phonology, and pragmatics. However, these areas of language are rarely affected to the same degree. In English, vocabulary and pragmatic skills are often relative strengths, whereas phonology and especially morphosyntax are relative weaknesses. This profile is not seen in all English-speaking children meeting the criteria for SLI. For example, some children with SLI show notable word-finding problems. There have been several attempts at determining whether the differences seen among children with SLI constitute distinct subtypes or instead represent different points on a continuum. Resolution of this issue will be important, as identification of the correct phenotype of SLI will be necessary for further progress in the genetic study of this disorder.

The heterogeneity of SLI notwithstanding, certain symptoms may have the potential to serve as “clinical markers” of SLI. For English-speaking children, two measures seem especially promising. One is a measure of the children's use of grammatical morphemes pertaining to grammatical tense and agreement, such as regular past -ed, third person singular -s, and copula and auxiliary forms of be (Rice and Wexler, 1996). The second is a measure of children's ability to repeat nonsense words containing several syllables (e.g., Dollaghan and Campbell, 1998). Both of these measures are quite accurate in distinguishing children with SLI from their normally developing age mates.

The linguistic profile of relative strengths and weaknesses in SLI seems to be shaped to a significant degree by the language being acquired. For example, children with SLI acquiring inflectionally rich languages such as Italian and Hebrew are not as severely impaired as their English-speaking counterparts in their use of grammatical inflections pertaining to tense and agreement. On the other hand, Swedish-speaking children with SLI show more serious problems in using appropriate word order than do children with SLI acquiring English (see Leonard, 1998, for a recent review).

Evidence for the efficacy of intervention is abundant in the literature. For example, for preschool-age children with SLI, approaches such as recasting have been relatively successful (Camarata and Nelson, 1992; Fey, Cleave, and Long, 1997). However, although the gains made in intervention usually go well beyond those that can be expected by maturation alone, no intervention approach has led to dramatic and rapid language gains by these children on a consistent basis. This is especially true when gains are defined in terms of use in spontaneous speech.

Attempts to explain the nature of SLI vary considerably. Most of these accounts focus on the extraordinary grammatical deficits often seen in children with SLI. It is possible to classify these alternative accounts according to their principal assumptions. Some accounts assume that children with SLI lack particular types of grammatical knowledge. For example, the “extended optional infinitive” account assumes that children with SLI go through a protracted period during which they assume that tense and agreement are optional rather than obligatory in main clauses (Rice and Wexler, 1996). The “representational deficit for dependent relationships” account assumes that children with SLI fail to grasp that movement or checking of grammatical features is obligatory (van der Lely, 1998).

Other types of accounts assume that children with SLI might have the potential to acquire normal grammar but have limitations in processing that slow their identification and interpretation of the relevant input and their ability to retrieve this information for production. In some cases the processing limitation is assumed to be quite general (Johnston, 1994; Ellis Weismer, 1996). In other cases the limitation is assumed to be specific to particular operations, such as phonological processing (Chiat, 2001) or the processing of brief or rapidly presented auditory information (e.g., Tallal et al., 1996).

Future research on SLI will make two types of contributions. Most obviously, greater understanding of this disorder should lead to more effective methods of assessment, treatment, and prevention. In addition, because the language disorder seen in SLI may in many cases occur in the absence of accompanying impairments, it constitutes a challenge for theories of language learning to explain.

See also language disorders in school-age children: overview; speech disorders in children: a psycholinguistic perspective.