aslha may 1 2009, volume 18, issue 2

7/30/2019 ASLHA May 1 2009, Volume 18, Issue 2

1/102

Change

Iwas recently asked to give aguest lecture in a colleaguesclass on the topic of language-

learning disabilities. To preparefor the lecture, I went back to oldcourse files on my computer fromwhen I taught the courseLanguageDisorders: School-Age Popula-tions back in 2001. To my sur-

prise, among my course files, therewere no PowerPoint slideshows for me to review and (hope-fully) borrow content fromrather, I found lecture notes

written in longhand that corresponded to a set of overheadtransparencies. It is almost impossible for me to recall teach-ing from lecture notes and transparencies, despite it only

being a relatively short time ago when apparently I did! Idont recall the transparency-to-PowerPoint transition as

being particularly difficult, although I do recall there being aperiod where I would bring overhead transparencies of myPowerPoint slides as a backup when giving conference

presentations. And, Ill admit, occasionally I feel nostalgic formy transparencies, particularly in those perennial instanceswhen technology lets me down.

My experience reflects a broader social shift as we tran-sition from relying on print to electronic media in manyfacets of life. Presently, I use online resources to pay my

bills, book hotel accommodations and airline travel, register

for conferences, researchrestaurant menus and movie schedules,print airline boarding passes, track my spending, sampleand purchase music, correspond with family and friends, andshop for books, clothing, and household merchandise. Justlast year, I finally transitioned from a paper calendar to anonline Outlook-based calendar; of all the transitions from

paper to electronic media, the shift to online scheduling wasthe most challenging for me (and the one I most stronglyresisted).

There is yet another impending print-to-electronic-mediatransition that will affect not only me but all the readers ofASHA journals, and that is the upcoming shift to online-only

journals scheduled to happen with the 2010 volumes. Thebenefits of going to online-only journals are numerous and

include cost-effectiveness and sensitivity to the environment,to name just a few. Currently, I suspect that many readersof this column already utilize ASHA resources available

online at www.asha.org, to include not only the compre-hensive archives of all previous issues ofAJSLPbut also avariety of other key archival documents (e.g., the online Leader,Code of Ethics, and cardinal documents of the Association).Importantly, the shift to online-only journals will allow in-tegration of new features and tools that could only be possible

by going online. Specifically, efforts are under way to offerenhanced online supplementary content linked to journalarticles, such as videos, sound clips, and images designed tohelp readers better understand content presented in articles.For instance, a research article involving presentation ofauditory stimuli to adult listeners could be coupled with

sound clips so that readers could hear the exact stimuli used inthe study. Because replication of study findings is critical tothe process of scientific accumulation, the possibilities pre-sented by attaching enhanced online content to articles are

particularly exciting.Over the next months, I encourage readers ofAJSLP to

begin to prepare for the transition from paper to online-onlyjournals so that it is as seamless as possible. You can do thisby familiarizing yourself with the available online versionof the current issue of the journal as well as the entire archiveat the journals home page (http://ajslp.asha.org/). Addition-ally, go ahead and follow the Sign Up for E-mail Alertslink on the AJSLPhome page so that you can receive thetable of contents directly to your inbox as each AJSLPissueis released. When you begin to receive these alerts, follow

the link and browse the journal online so that you becomeaccustomed to examining the online version of the journalrather than the print version, which in the near future will nolonger be available.

I recognize that change is difficult, yet it is also the path-way to possibility. While many of us will lament the lossof our print journals for some time once they go away, Isuspect there will be a time not far into the future when wereflect on those old print journals in the same way that Ireflected on the overhead transparencies with which I usedto lecture: slightly nostalgic but glad we moved on.

Laura JusticeEditor

From the Editor

114 American Journal of Speech-Language Pathology Vol. 18 May 2009 DOI: 10.1044/1058-0360(2009/ed-02)


2/102


3/102

Is Expressive Language Disorder an Accurate

Diagnostic Category?

Laurence B. Leonard

Purdue University, West Lafayette, IN

Purpose: To propose that the diagnostic category ofexpressive language disorder as distinct from a disorderof both expressive and receptive language might not beaccurate.Method: Evidence that casts doubt on a pure form of thisdisorder is reviewed from several sources, including theliterature on genetic findings, theories of language impair-ments, and the outcomes of late talkers with expressivelanguage delays. Areas of language that are problematicin production but not readily amenable to comprehensiontesting are also discussed.Conclusions: The notion of expressive language disorderhas been formalized in classification systems and is implicitif not explicit in the organization of many standardized tests.However, a close inspection of the evidence suggests thatdeficits in language expression are typically accompanied bylimitations in language knowledge or difficulties processinglanguage input. For this reason, the diagnostic category ofexpressive language disorder should be used with consider-able caution. This view has implications for both research andclinical practice.

Key Words: expressive language disorder,specific language impairment, language disorders

In the literature on children with language impairments,it is common to find reference to the heterogeneity of this

population. Although some patterns of strengths andweaknesses are more common than others, exceptions to thecommon patterns can easily be found. These differencesamong children have encouraged attempts to identify sub-groups of children with language impairments (Aram &

Nation, 1975; Conti-Ramsden, Crutchley, & Botting, 1997;Korkman & Hkkinen-Rihu, 1994; Rapin, 1996; Rapin &Allen, 1983, 1987; van Daal, Verhoeven, & van Balkom,2004; Wilson & Risucci, 1986; Wolfus, Moscovitch, &Kinsbourne, 1980). Often these subdivisions reflect domainsof language (e.g., grammar, vocabulary) that may be es-

pecially weak in some but not other subgroups. However,just as frequently, subdivisions are based on the modalitycomprehension and/or productionthat may be affected.This modality distinction is most often a division betweenchildren with expressive language deficits and childrenwith receptive-expressive language deficits. Children in

the first category have problems that are principally confinedto language output; children in the latter category exhibitsignificant weaknesses in language comprehension as wellas language expression.

The distinction between expressive language disorderand receptive-expressive language disorder is not simply aninformal clinical sorting of children; it has been formalized.Expressive language disorder and (mixed) receptive-expressive language disorder constitute categories in the

Diagnostic and Statistical Manual of Mental Disorders(4th ed., text revision; American Psychiatric Association,2000) and carry different codes (315.31 and 315.32 for theexpressive and receptive-expressive forms of the disorder,respectively). In the International Statistical Classificationof Diseases and Related Health Problems, Tenth Revision(ICD-10; World Health Organization, 2005), the two sub-types are referred to as expressive language disorder andreceptive language disorder (with codes of F80.1 and F80.2,respectively). Although the latter term does not includethe word expressive, the ICD-10 definition of this subtypespecifies thatin almost all cases expressive language ismarkedly disturbed (p. 238). To avoid confusion, the termreceptive-expressive language disorder will be used here.

An inspection of standardized language tests can givethe impression that the distinction between expressive lan-guage disorder and receptive-expressive disorder should be

part of a clinicians a priori assessment strategy. Many testshave a separate score for expressive and receptive language,and those tests with multiple subtests have provisions forcombining scores of particular subtests to arrive at an ex-

pressive language composite score and a receptive languagecomposite score. Tests that provide separate expressiveand receptive scores include the Clinical Evaluation of Lan-guage FundamentalsPreschool, Second Edition (Wiig,Secord, & Semel, 2004), the Clinical Evaluation of Lan-guage Fundamentals, Fourth Edition (Semel, Wiig, & Secord,2003), the Comprehensive Assessment of Spoken Language(Carrow-Woolfolk, 1999), the Oral and Written Language

Scales (Carrow-Woolfolk, 1995), the Preschool LanguageScale, Fourth Edition (Zimmerman, Steiner, & Pond, 2002),the Reynell Developmental Language Scales (U.S. edition;Reynell & Gruber, 1990), the Sequenced Inventory of Com-munication Development, Revised Edition (Hedrick, Prather,& Tobin, 1995), the Test of Adolescent and Adult Language,Third Edition (Hammill, Brown, Larsen, & Wiederholt, 1994),the Test of Early Language Development, Third Edition(Hresko, Reid, & Hammill, 1999), the Test of LanguageDevelopmentIntermediate, Fourth Edition (Hammill &

Newcomer, 2008), the Test of Language Devel opmentPrimary, Fourth Edition (Newcomer & Hammill, 2008),

Viewpoint

American Journal of Speech-Language Pathology Vol. 18 115123 May 2009 A American Speech-Language-Hearing Association

1058-0360/09/1802-0115

115


4/102

and the Test of Narrative Language (Gillam & Pearson,2004).

The distinction between expressive and receptive lan-guage has also become highly relevant to evidence-based

practice. For example, Law, Garrett, and Nye (2004) made

use of this distinction in their meta-analysis of interventionstudies meeting the criteria for randomized controlled trials.One of their conclusions from their secondary analyses wasthat there may be a differential effect of intervention forexpressive syntax, with intervention being effective for thosechildren who do not also have receptive language difficul-ties (p. 931).

In this article, I point out significant limitations in theexpressive versus receptive-expressive dichotomy as appliedto vocabulary, grammar, and narrative skills. I recognize thatchildren with weak expressive language ability in these areascan vary widely in their language comprehension ability. Ialso acknowledge that, using psychometric criteria, it is often

possible to group children into expressive and receptive-

expressive categories. However, I question whether, at adeeper level, the distinction is accurate.

A simple example can be used to introduce this idea.Mainstream American English-speaking children with lan-guage impairments often produce not in contexts requiringdoesnt(e.g., Mommy not like carrots). It is highly likely thatchildren making this error in production can understandsentences such as Mommy doesnt like carrots, as the formdoesntwill have been heard before in similar contexts. (Onecan imagine that, in response to Show me Mommy doesntlike carrots, children might select a picture of a womanwho is frowning while looking at a plate of carrots ratherthan a picture that depicts a woman smiling at the carrots.)I would argue that the childrens failure to produce thecorrect form is due to insufficient knowledge; the childrenmay have sufficient familiarity with doesntand the contextsin which it appears to interpret its meaning, but not knowthis form well enough to recognize that it should be retrievedfor use in their own speech.

It is important to stress that such children might wellearn age-appropriate scores on sentence comprehension testsand low scores on sentence production tests. This patternof performance might give us license in a technical senseto say that these children have an expressive languagedisorder. However, this is not the same as saying that their

problems in using forms such as doesnt are limited to out-put. Insufficient knowledge is also a factor. In making mycase, I will try to keep separate the concept of psychomet-rically defined gaps between expressive and receptive scores

on the one hand and the concept of limitations in languageknowledge on the other. Children showing gaps betweenscores on expressive and receptive language tests might bedescribed differently from children showing low scores on

both types of tests, but it seems risky to describe the out-put problems of the first group as a limitation in expressivelanguage only. I will pursue these issues more fully below.I begin with some of the pitfalls in defining expressivelanguage disorder based on differences between expressiveand receptive language test scores, and then follow witha more extended discussion of how expressive languagedisorder may be an inaccurate characterization because the

degree or type of language knowledge needed for languageexpression may differ from that needed to succeed in lan-guage comprehension.

Empirical Discrepancies and Gaps in the DataThere is little doubt that early methods of defining chil-

dren as exhibiting an expressive or receptive-expressivelanguage disorder were fraught with problems. Bishop(1979) pointed out some of these difficulties in a now-classicstudy. She administered both the Peabody Picture Vocabu-lary Test (PPVT; Dunn, 1965) and the Test for Receptionof Grammar (Bishop, 1977) to children who had been clas-sified as showing either an expressive language disorder or areceptive-expressive language disorder. Bishop found that

both groups of children scored well below the level of age-matched typically developing children on both tests. Thetypically developing children in this study were not unrep-resentative; their standard scores averaged approximately

100. Given the lower scores of the expressive languagedisorder group on two different comprehension measures,Bishop argued that there was no clear justification for givingthese children a clinical label that excluded reference toreceptive language.

We now have greater sophistication in using test scores asa basis for classifying children as exhibiting an expressivelanguage disorder or a receptive-expressive language disor-der. Factors now considered include the need to select onlythose expressive and receptive language tests that have ac-ceptable levels of sensitivity and specificity, and the need toensure that the standard errors of measurement of the expres-sive and receptive language tests are taken into account

before concluding that the two types of scores are trulydifferent.

Of course, it is also important to take into account thedomains of language that are assessed. For example, Deevyand Leonard (2004) studied a group of children with lan-guage impairments who earned low-average scores on thePPVT, Third Edition (PPVTIII; Dunn & Dunn, 1997), areceptive vocabulary test, but scored poorly on tests ofexpressive language. When tested for their understandingof wh-questions of the type Who was the happy brown dogchasing? the children performed significantly below thelevel of a group of slightly younger typically developingchildren who were matched according to raw scores onthe PPVTIII. Even though these childrens low-averagereceptive vocabulary scores were not clinically significant,their poorer understanding of wh-questions relative to youn-

ger typically developing children would seem to rendera classification ofexpressive language disorder quiteinsufficient.

Perhaps the greatest obstacle to using test scores to clas-sify children as showing an expressive language disorderrests in the fact that some of the details of expressive lan-guage that are most problematic for children with languageimpairmentsand are often of great diagnostic importanceare extremely difficult to test in comprehension. A clearexample is seen in the area of grammatical morphology.In production, morphemes that reflect tense and agree-ment, such as auxiliary is, third person singulars, and

116 American Journal of Speech-Language Pathology Vol. 18 115123 May 2009


5/102

past tense ed, often pose significant problems for childrenwith language impairments. These children often use thesemorphemes inconsistently for an extended period. Thisinconsistency represents an especially good means of dis-tinguishing children with language impairments from their

typically developing peers; sensitivity and specificity valuesapproximate or exceed 80% (Bedore & Leonard, 1998; Rice& Wexler, 2001). However, tests of the comprehension ofthese morphemes are difficult to develop, for several reasons.First, given the structure of English, verb morphemes thatreflect subject-verb agreement (e.g., auxiliary and copula beforms, third person singulars) are accompanied by thesubject of the sentence that provides a supporting (or, in thecase of testing, a confounding) cue. To control for this fact,items with invariant nouns must be employed. Unfortunately,children may not know that nouns such as fish, deer, andmoose are identical in singular and plural. Thus an item ona picture-pointing task such as Show me The fish are swim-ming may be difficult either because the child does not

understand that auxiliary are marks third person plural orbecause the child does not know thatfish is the form usedfor plural as well as singular. (In contrast, in languages suchas Spanish and Italian in which the subject is optional whenthe referent is clear, the item can be presented without thesubject, thus avoiding subject number cues and the needfor invariant nouns.)

Johnson, deVilliers, and Seymour (2005) avoided thisproblem in a picture-pointing task with the clever innovationof using verbs with word-initial /s/. By ensuring that therewere no pauses between the subject and the verb, contrastssuch as The duck swims in the water and The ducks swimin the water could be used without providing confoundingcues. Unfortunately, typically developing children do notappear to perform above the level of chance until 5 years ofage on this type of task. As noted by Johnson et al., typicallydeveloping children produce third person singulars to agreater degree than their performance on this task would

predict. Therefore, the type of knowledge required for pro-duction must be somewhat different than the knowledgerequired to succeed on this task.

As children reach 4 years of age, the assessment oftheir understanding of certain agreement morphemes is

possible through the use of grammaticality judgment tasks.In the Rice/Wexler Test of Early Grammatical Impairment(Rice & Wexler, 2001), childrens judgments of sentenceswith missing agreement morphemes (e.g., He running awayor Now the bear want a drink) and agreement morphemesreflecting incorrect agreement (e.g., He are mad or I drinks

milk) are evaluated, along with sentences possessing correctuse of these morphemes. The advantage of this assessmenttool is that a separate evaluation can be made of the chil-drens judgments of sentences most likely to resembletheir own production errors (notably, the omission ofagreement markers) and their judgments of sentenceswhose errors are not those likely to be used by the children(those with overt agreement errors such as I drinks milk).However, because grammaticality judgment tasks requiresome degree of metalinguistic skill, high levels of bothsensitivity and specificity are not seen until approximately6 years of age.

The assessment of childrens comprehension of pasttense poses a different type of problem. When assessed

by means of a picture-pointing task, past tense is typicallydistinguished from present tense by contrasting a drawing ofan action that was just completed with a drawing of an ac-

tion still in progress. For example, the drawing for an itemassessing comprehension of The girl jumped could depicta girl landing after having just jumped over a fence, con-trasted with a drawing of a girl still in the air. However, sucha past tense item conflates past tense with completion or

perfective aspect. Within a picture-pointing format, it wouldbe very difficult to test for past tense without providing cuesof this type, yet the absence of such cues is necessary todetermine whether the child understands past tense inde-

pendent of perfective aspect.I suspect that our limited ability to assess certain language

details in comprehensionor the older ages at which wehave had to assess themhas contributed to the impres-sion that some children have deficits restricted to language

expression. Quite possibly, this complication has been re-sponsible in part for the apparent instability of this diagnosticcategory. For example, using the categories of expressiveand receptive-expressive language deficits, Conti-Ramsdenand Botting (1999) found that many of the children whowere classified as exhibiting an expressive language disorderwere reclassified as showing a receptive-expressive languagedisorder when tested 12 months later. Tomblin and Zhang(2006) tested alternative models for their suitability in ac-counting for childrens scores on standardized language test

batteries at four different ages. They found that a single-dimension modelin which all language tests were treatedas a single factorwas superior to a model that treated ex-

pressive and receptive scores as separate factors. There wassome evidence that, across time, grammatical abilities andvocabulary abilities became differentiated, but an expressiveversus receptive distinction did not emerge.

Limitations in Language Knowledge

Underlying Seemingly Pure Cases

of Expressive Language Disorder

With further development and refinement of testingprocedures, it is likely that we will move toward a greaterunderstanding of childrens language comprehension. How-ever, I believe that problems in expressive language may

be due in part to limitations in the degree of childrens lan-guage knowledge, and this graded level of knowledge may

prove very difficult to measure given current methods oftesting in which a response to any receptive test item is eithercorrect or incorrect. The subtle but important role played

by language knowledge can be seen if we consider thoseoperations that are most often associated with expressivelanguage disorders.

These operations include those involved in retrieving andpreparing linguistic material for output. (The problematicoperations cannot be confined to the physical act of speaking.If this were the case, all children with expressive languagedisorders would have significant phonological difficultiesand there would be no way to distinguish problems in a

Leonard: Expressive Language Disorder 117


6/102

domain such as grammar from those in a domain such asvocabulary.) I believe that the difficulties that children withexpressive language disorder have with operations of re-trieval and preparation for output are exacerbated by limi-tations in language knowledge. I consider the operations of

word retrieval and sentence formulation in turn.

Word Retrieval Problems

Word retrieval difficulty is usually defined as a problemin accessing words that are already known by the child.However, the word-finding literature offers an alternativeinterpretation. McGregor, Newman, Reilly, and Capone(2002) provided an illustrative study in this regard. Theseinvestigators performed a comprehensive examination ofthe semantic knowledge possessed by a group of childrenwith specific language impairment (SLI) who were foundto commit a substantial number of naming errors for age-appropriate objects. They found that the childrens drawings,definitions, and recognition responses were also relatively

poor for objects that were incorrectly or inadequately named.Most of the children earned age-appropriate scores on thePPVTIII. McGregor et al. summarized their findings suc-cinctly: This study demonstrates that the degree of knowl-edge represented in the childs semantic lexicon makeswords more or less vulnerable to retrieval failure and thatlimited semantic knowledge contributes to the frequentnaming errors of children with SLI (p. 998).

Limitations in degree of word knowledge can also affectnaming response time (RT), even when children producethe correct name for an object (e.g., Kail & Leonard, 1986).The typical explanation for this view is that repeated en-counters with a word lead to stronger and more numerousassociations in semantic memory. Words with stronger and

more numerous associations can be retrieved more quicklythan words that are represented in semantic memory withfewer and weaker associations. For this reason, typicallyfunctioning adults show faster RTs for names that have highfrequency of occurrence in the language than for names withlower frequency of occurrence. Obviously, these adults donot have selective retrieval deficits; rather, the RT differ-ences reflect differences in the degree to which the high-and low-frequency words are known. It would follow thatthe slower RTs for children with language impairments thathave been reported in some picture-naming studies mightwell be attributable to limitations in the degree to which thechildren knew the words.

Sentence Formulation Problems

Another possible deficit of a strictly expressive natureis a problem of sentence formulation, that is, a deficit in

preparing already-acquired language material into sentencesfor output. One relevant line of evidence is the study ofspeech disruptions in children with language impairments.If children insert pauses or fillers (e.g., uh or well) or re-

peat syllables or words in the sentences they produce,they may be having difficulties with sentence formulation,even when the sentences contain no grammatical errors.Finneran, Leonard, and Miller (in press) found that a group

of 9-year-olds with SLI produced grammatical sentenceswith significantly more speech disruptions than a groupof same-age peers. Similar results were reported by Guo,Tomblin, and Samelson (2008), who found that children withSLI had a significantly higher number of pauses than same-

age peers in their production of narratives. The fact thatthe rate of pauses was higher at phrase boundaries led theseinvestigators to conclude that these pauses may have beendue to lexical and/or syntactic weaknesses in the childrenwith SLI. As we saw in the discussion of word retrieval

problems, speech disruptions occurring during sentenceproduction could reflect words or syntactic structures thatare simply not as well known by children with languageimpairments as by typically developing peers, thus requiringmore of a struggle to accurately retrieve. The source of thedifficulty, then, may occur prior to the point of preparing theutterance for production.

Another look at sentence formulation is provided bystudies that employ syntactic priming. Leonard et al. (2000)

found that children with language impairments who wereinconsistent in using auxiliary is were more likely to describea target picture with this morpheme (e.g., The Grinch isreading the book) if they had just repeated a prime sentencesuch as The cats are drinking the milk than if they had justrepeated a sentence such as The bird flew away. The dif-ference between these two priming conditions was greaterin the group of children with language impairments than ina group of younger typically developing children who werealso inconsistent in their use of auxiliary is. The primingeffects seen in both groups were interpreted as reflectingthe prior activation of a syntactic frame. In the case of Thecats are drinking the milk, the frame is appropriate for usewhen describing the target picture, and its prior activationrenders it easier to retrieve. Once retrieved, the content wordsand function words (including the specific auxiliary form, is)can be retrieved and inserted into the frame. The fact thatthe priming effects were larger in the group with languageimpairments led Leonard et al. to propose that these childrenhad greater difficulty with sentence formulation, and this

process was greatly assisted through prior activation of anapplicable syntactic frame. Similar results were obtained ina subsequent study by Leonard et al. (2002).

Although formulation of an utterance seems to fall on theexpressive side of language, it is not an insulated process.More recent work provides a strong indication that primingcrosses modalities. In children (Shimpi, Gmez, Huttenlocher,& Vasilyeva, 2007) as well as adults (Branigan, Pickering,Stewart, & McLean, 2000), simply hearing prime sentences

without repeating them also leads to increased use of thesyntactic frame in production. The fact that production isinfluenced by prime sentences that are heard but not repeatedhas led to the view that language production relies on thesame type of structural knowledge as language comprehen-sion (Bock, Dell, Chang, & Onishi, 2007). It would followthat, in the Leonard et al. studies, the prime sentences fa-cilitated the childrens knowledge of the sentence structure,not just their ability to call on the structure for use in a targetsentence.

It can be seen, then, that the findings from word retrievaland sentence formulation studies do not provide sufficient



7/102

evidence to conclude that the problem lies exclusively inlanguage expression. Limitations in language knowledge arealso implicated.

The Expressive Versus Receptive-ExpressiveDistinction in the Identification, Prediction,

and Explanation of Language Impairments

There are additional reasons to question the notion ofexpressive language disorder. An inspection of the literatureon the identification of language impairments, the predictionof later language impairments, and theories of languageimpairments provides very little (if any) evidence that ex-

pressive language problems occur in isolation. Instead, theyseem to be accompanied by weaknesses in language com-

prehension and/or knowledge.

Genetic Studies and the Expressive Versus

Receptive-Expressive Distinction

Twin studies have provided valuable information con-cerning the genetic and environmental influences on chil-drens language abilities. Bishop, Adams, and Norbury(2006) employed the twin-study methodology and identifiedtwo heritable weaknesses associated with risk for languageimpairment. One was a limitation seen on a nonword rep-etition task. The other was a weakness in grammaticalcomputation, as reflected on tasks of tense and agreementmorpheme production and syntactic comprehension. Al-though heritable, these two weaknesses were separable;one could occur without the other. Neither of these weakability areas suggests a division between receptive andexpressive skills. The grammatical computation measureinvolved tasks of both production and comprehension. Thenonword repetition task, although requiring a productionresponse, is often taken to be a measure of verbal short-termmemory (e.g., Gathercole & Baddeley, 1990), and thuscannot be viewed as a purely expressive task.

Theories of Grammatical Impairment and the

Expressive Versus Receptive-Expressive Distinction

There are several prominent explanations for the typesof grammatical difficulties experienced by children withlanguage impairments. However, all of these accounts seemto assume that the problem is not limited to language ex-

pression but extends to the childrens incomplete grasp of

particular linguistic principles, or to their inability to processlinguistic information in the input. Although there is debateamong researchers about the descriptive and explanatoryadequacy of some of these accounts, most of these accountswould have been dealt a fatal blow if a significant minorityof the children serving as participants in these studies per-formed adequately on the comprehension or receptivelanguage-processing tasks that were used to test these ac-counts. For example, the extended optional infinitive accountof Rice, Wexler, and their colleagues holds that childrenwith language impairments fail to grasp the notion that tenseand agreement are obligatory in main clauses (Rice, 2003;

Rice & Wexler, 1996; Rice, Wexler, & Hershberger, 1998).Instead, they treat tense and agreement as optional. In pro-duction, this problem leads to inconsistency in the use oftense and agreement morphemes. In comprehension, it isseen when the children judge sentences such as The boy

am running as wrong but treat both The boy is running andThe boy running as acceptable. Similarly, in the Represen-tational Deficit for Dependent Relationships account ofvan der Lely, children with language impairments havedifficulty not only in the use of certain grammatical de-tails but also in comprehending them (van der Lely, 1998;van der Lely & Battell, 2003). Thus wh-questions that re-quire movement ofwh-words and auxiliary verbs, as in Whowas the girl kissing? are more difficult for these childrenin both comprehension and production than wh-questionsthat can be produced or interpreted with no such movement,as in Who was kissing the girl? Recent approaches thatdescribe the movement deficit somewhat differently never-theless find that comprehension is affected (Friedmann &

Novogrodsky, 2007).Ullman and Pierpoint (2005) propose that many children

with SLI have a deficit in the neural circuitry responsiblefor procedural memory, the system involved in the learningand execution of sequential cognitive (including linguistic)information. This procedural deficit is assumed to affectcomprehension as well as production. Although more re-search is needed to test this proposal, recent evidence on

procedural learning difficulties in children with languageimpairments indicates that the problem is not limited to lan-guage expression (Tomblin, Mainela-Arnold, & Zhang, 2007).

Accounts that assume processing capacity limitations inchildren with language impairments also implicate compre-hension as well as production. Put more precisely, theseaccounts assume that the problem of these children rests in alimited processing capacity that restricts the amount andtimeliness of information that can be taken in, thus impedingthe development of comprehension as well as production.Perhaps the dominant proposal of this type is that thesechildren have significant limitations in verbal workingmemory (e.g., Hoffman & Gillam, 2004; Leonard et al.,2007; Montgomery, 2000). Studies that have examinedverbal working memory in children with language impair-ments consistently report difficulties in this population.

It is tempting to treat differences between linguisticallybased accounts and those that assume processing limitationsas equivalent to a difference in competence versus perfor-mance. A limitation in competence can easily be viewed as areceptive-expressive problem given that language knowl-

edge is affected. However, a processing limitation affectsmore than online performance; if the information is notadequately processed, it cannot serve to form or strengthenan underlying representation in the childs developing lan-guage system. As a result, knowledge is affected, not simplythe childs performance in the moment.

It is striking that there are no theories of expressive (only)language disorder apart from proposals that pertain to seg-mental phonology or prosody (e.g., Gerken & McGregor,1998; Goffman, 2004). Despite the fact that the expressivecomponent has been front and center in the existing theoriesof grammatical impairments, the proponents of all theories



8/102

have seen the need to assume deficits that extend beyondlanguage output.

Late Talkers and the Prediction of Outcomes

Several excellent prospective studies have been con-ducted that follow late talkers for several years with theaim of determining which factors represent risk factors forlanguage impairment at a later age. Children are definedas late talkers according to expressive language criteria,specifically, as producing fewer than 50 words and no wordcombinations (e.g., Rescorla, 1989, 2005), or as falling

below the 10th percentile in word use at 24 to 30 months(e.g., Ellis Weismer, 2007; Thal, 2005; Thal, Tobias, &Morrison, 1991). Many of the late talkers in these studiesearned age-appropriate scores on tests of language compre-hension. A careful review of this literature reveals one

puzzling fact. The percentage of late talkers with outcomesthat lead to a diagnosis of SLI at a later age is consistently

much lower than would be expected given that the preva-lence of SLI is approximately 7% at 5 years of age (Tomblinet al., 1997). For example, Thal (2005) found that only 8.8%of the late talkers in her prospective study met the criteriafor SLI when the children were 5 years old. Ellis Weismer(2007) found that only 7.5% of the late talkers in her inves-tigation met the criteria for SLI when they reached age 5.These percentages are close to the prevalence figures forSLI among 5-year-olds in the general population and suggestthat an early pattern of slow expressive language develop-ment is not a good predictor of later language impairment.Instead, early, slow expressive language development seemsto predict later language functioning that is below average

but within normal limits (e.g., Rescorla, 2005). Importantly,when 24-month-olds are found to have low comprehen-

sion as well as production ability, their outcomes are poorer(Thal et al., 1991). Ellis Weismer (2007) reported that latetalkers comprehension at 30 months was the strongest sin-gle predictor of these childrens language production scoresat 66 months.

In a more recent investigation, Rice, Taylor, and Zubrick(2008) assessed the language abilities of 7-year-olds whohad been identified as exhibiting either late language emer-gence or language emergence at a typical age. Late languageemergence was defined as a small expressive vocabulary(70 words or fewer) or no word combinations at 24 monthsof age. At age 7 years, a higher percentage of the late lan-guage emergence group met the criterion for affectednesson 7 of the 17 language measures obtained. The percent-

age of children in this group who met the criterion rangedfrom 4 to 23, depending on the language measure. Thesepercentages are somewhat higher than those of previousstudies. However, it should be noted that the criterion usedfor affectedness for each language measure was 1 SD. Sucha criterion is not especially stringent, as it represents thelowest 16% of a distribution.

Implications for the Study of Heterogeneity

If a conventional distinction such as the expressive versusreceptive-expressive distinction becomes suspect, can other

conventional distinctions be far behind? For example, eversince the influential work of Bloom and Lahey (1978), thedistinction among content, form, and use (including areasof overlap) has been viewed as important for the descriptionof language impairments in children, even making its way

into the definition oflanguage disorder by the AmericanSpeech-Language-Hearing Association (1993). This dis-tinction has served as a useful heuristic, by focusing ourattention on broad dimensions of language that might beadversely affected. However, to my knowledge, it has notyet been demonstrated that children can be reliably placedinto subtypes that conform to these particular divisions. Thedegree to which these and other conventional distinctionscan be substantiated would seem to be an important topicfor future investigation.

It is possible that meaningful subtypes might be identifiedthrough genetic studies of potential endophenotypes (clus-ters of related abilities) that arise from theoretical propos-als of causal factors in language impairment (Bishop, 2006).

For example, if weaknesses in several theoretically relatedabilities appear to cluster together in monozygotic twins toa greater extent than in dizygotic twins, this cluster mightconstitute a meaningful subtype of language impairment.Additional clusters might also be discovered in this way, andthese might prove to be genetically distinct from each other.Still other weaknesses may prove to be rather frequent butunreliable in their patterning. The latter weaknesses might

be regarded as secondary deficits that may accompany oneor more of the core subtypes but not enter into a formalclassification scheme.

Implications for Clinical Research and Practice

One goal for future clinical research would be to developor refine methods for assessing childrens understandingof language details that are often problematic for these chil-dren in production but have proven to be difficult to assess incomprehension. As noted earlier, one such language detailis subject-verb agreement. Current methods are suitable forages 5 years and older, but less than ideal for younger chil-dren. However, it might be possible to assess childrenssensitivity to subject-verb agreement at a younger age.McNamara, Carter, McIntosh, and Gerken (1998) found that

preschool-age children with SLI were more likely to pointto a correct picture (e.g., a picture of a bird) in response toa sentence containing an appropriately used article (e.g.,

Find the bird for me) than in response to a sentence con-taining an inappropriate morpheme (e.g., Find was bird for

me). Sentence contexts appropriate for a subject-verb agree-ment morpheme were not employed (e.g., contrasting Momis running with Mom the running), but it seems that this taskmight allow for items of this type. Of course, as I argued ear-lier, an awareness that a morpheme seems to be in an ap-

propriate (or inappropriate) context is no assurance that thechild has sufficient knowledge to use the morpheme. Addi-tional forms of assessing childrens receptive command ofsubject-verb agreement will still be needed.

The assessment of childrens comprehension of past tensehas also proven to be challenging. However, it may be pos-sible to develop informative tests of past tense comprehension



9/102

by adapting existing procedures seen in the normal childlanguage literature. Wagner (2001) developed a task used withyoung typically developing children in which a toy characteris made to proceed along a path, performing actions in themiddle of the path and then again at the end of the path. The

actions performed in the first location can have definableendpoints (e.g., filling a toy bucket with small toy apples).In some items, this action might be completed before thecharacter moves on to the next location (e.g., all the applesare placed in the bucket), and in other items, the action mightnot be completed before the character resumes the journeyalong the path (e.g., some of the apples remain on the groundnext to the bucket). The action at the end of the path is iden-tical to the first action. While the character is performingthe action at the end of the path, the examiner can ask the childto point to the location of the named action. A request suchas Show me where the girl filled up the bucket is most appro-

priate when the first action had been completed (all appleswere placed in the bucket). A request such as Show me where

the girl was filling up a bucketcan be made both when the firstaction had been completed and when it was left incomplete,

because the past progressive does not entail completion,only past time. To contrast past with present tense, the requestShow me where the girl is filling up a bucket can be made.In this instance, of course, the correct response is the locationof the action being performed at the end of the path.

The views expressed in this article also carry implicationsfor clinical practice. When children show a significant gap

between their expressive and receptive language test scores,clinicians should carefully examine the details of expressivelanguage that are problematic for the children. If these detailsare not reflected adequately in the receptive language teststhat were administered, there is a possibility that the childrenlack the prerequisite knowledge for successful production.If receptive measures that do tap into these details are avail-able, clinicians might then employ them to supplement thereceptive language testing of the children.

Treatment decisions, too, might be influenced by theviews conveyed here. Treatment focused on details ofexpressive language for which the childrens prerequisiteknowledge has not been established should involve anapproach that provides the children with information aboutmeaning, grammatical function, and/or contexts of use alongwith any production practice that is provided. For youngchildren, of course, such information might have to be

provided through examples rather than through formalinstruction.

I think it rather paradoxicaland consistent with the

views expressed here

that approaches such as recastinghave proven promising as a method for facilitating youngchildrens expressive language (e.g., Camarata & Nelson,2006; Camarata, Nelson, & Camarata, 1994; Leonard,Camarata, Pawowska, Brown, & Camarata, 2008). In thisapproach, clinicians respond to childrens utterances withconversationally appropriate utterances that resemble thechildrens preceding utterances but contain the languagetarget. Such an approach provides children with contextualinformation and contrasts between their own utterance andthe recast utterance, yet no expressive use of the target iseven required of the children. Of course, future research may

reveal that some children require practice in producing thelanguage target. However, I suspect that any treatment approachthat is found to meet the highest levels of evidence will havea significant component devoted to providing children withinformation that goes well beyond the act of production.

Conclusions

At the outset of this article, I introduced the idea thatchildren may exhibit deficits in expressive language that arecaused by limitations in knowledge that extend beyond a

problem with the retrieval and preparation of language ma-terial for output. These limitations might not be reflectedin the childrens scores on language comprehension tests. AsI have tried to show, current methods of assessing compre-hension abilities do not yet allow us to test childrens graspof certain details of language that are known to be prob-lematic in language expression.

It is also difficult to demonstrate that a pure expressive

language disorder can even exist, given the types of languageknowledge that seem to underlie operations such as wordretrieval and sentence formulation. I am also struck with howlittle the notion of expressive language disorder enters intoattempts to explain grammatical impairments. Outside ofthe realm of segmental phonology and prosody, there seemsto be no theory of expressive language problems that doesnot also assume a limitation in language knowledge or a prob-lem in processing language input. Furthermore, it does notseem plausible that the existing theories would remain viableif many of the children with language impairments in thesestudies could succeed in the comprehension and receptivelanguage-processing tasks that were employed to test thesetheories. Genetic evidence, too, seems to point to weaknessesin ability areas that incorporate receptive as well as expressivelanguage. Finally, early delays in expressive language (only)do not serve as good predictors of later language impairment;on the other hand, if comprehension delays are also seen at ayoung age, later problems in language are more likely.

It may be that the term expressive language disorderis useful in particular circumstances, as a type of shorthandto refer to children whose receptive language test scoresare demonstrably higher than their expressive language testscores. However, considering the questions that remain aboutthis diagnostic category, we should be alert to the possibilitythat children may lack the knowledge needed to producelanguage adequately even when their receptive languagescores might suggest otherwise.

Acknowledgments

This work was supported in part by National Institute onDeafness and Other Communication Disorders Grant R01 DC00458.Many thanks to Patricia Deevy and Jeanette S. Leonard for theirvery helpful comments on earlier versions of this article.

References

American Psychiatric Association. (2000). Diagnostic andstatistical manual of mental disorders (4th ed., text revision).Washington, DC: Author.



10/102

American Speech-Language-Hearing Association. (1993). Defi-nitions of communication disorders and variations[Relevant paper]. Available from www.asha.org/policy.

Aram, D., & Nation, J. (1975). Patterns of language behavior inchildren with developmental language disorders. Journal of

Speech and Hearing Research, 18, 229

241.Bedore, L., & Leonard, L. (1998). Specific language impairmentand grammatical morphology: A discriminant function analysis.

Journal of Speech, Language, and Hearing Research, 41,11851192.

Bishop, D. V. M. (1977). Comprehension of grammar: Normal andabnormal development. Doctoral thesis, Oxford University,Oxford, England.

Bishop, D. V. M. (1979). Comprehension in developmental lan-guage disorders. Developmental Medicine and Child Neurology,21, 225238.

Bishop, D. V. M. (2006). What cause specific language impair-ment? Current Directions in Psychological Science, 15,217221.

Bishop, D. V. M., Adams, C., & Norbury, C. F. (2006). Distinctgenetic influences on grammar and phonological short-term

memory deficits: Evidence from 6-year-old twins. Genes, Brainand Behavior, 5, 158169.

Bloom, L., & Lahey, M. (1978). Language development and lan-guage disorders. New York: Wiley.

Bock, K., Dell, G., Chang, F., & Onishi, K. (2007). Persistentstructural priming from language comprehension to language

production. Cognition, 104, 437458.Branigan, H., Pickering, M., Stewart, J., & McLean, J. (2000).

Syntactic priming in spoken production: Linguistic and temporalinterference. Memory and Cognition, 28, 12971302.

Camarata, S., & Nelson, K. E. (2006). Conversational recastintervention with preschool and older children. In R. McCauley& M. Fey (Eds.), Treatment of language disorders in children(pp. 237264). Baltimore: Brookes.

Camarata, S., Nelson, K. E., & Camarata, M. (1994). Com-parison of conversational recasting and imitative procedures for

training grammatical structures in children with specific lan-guage impairment. Journal of Speech and Hearing Research,37, 14141423.

Carrow-Woolfolk, E. (1995). Oral and Written Language Scales.Circle Pines, MN: AGS.

Carrow-Woolfolk, E. (1999). Comprehensive Assessment of Spoken Language. Circle Pines, MN: AGS.

Conti-Ramsden, G., & Botting, N. (1999). Classification ofchildren with specific language impairment: Longitudinalconsiderations. Journal of Speech, Language, and Hearing

Research, 42, 11951204.Conti-Ramsden, G., Crutchley, A., & Botting, N. (1997). The

extent to which psychometric tests differentiate subgroups ofchildren with SLI. Journal of Speech, Language, and Hearing

Research, 40, 765777.Deevy, P., & Leonard, L. (2004). The comprehension of wh-

questions in children with specific language impairment. Jour-nal of Speech, Language, and Hearing Research, 47, 802815.

Dunn, L. (1965). Expanded Manual for the Peabody PictureVocabulary Test. Circle Pines, MN: AGS.

Dunn, L., & Dunn, L. (1997). Peabody Picture Vocabulary TestThird Edition. Circle Pines, MN: AGS.

Ellis Weismer, S. (2007). Typical talkers, late talkers, and childrenwith specific language impairment: A language endowmentspectrum? In R. Paul (Ed.), Language disorders from a devel-opmental perspective (pp. 83101). Mahwah, NJ: Erlbaum.

Finneran, D., Leonard, L., & Miller, C. (in press). Speechdisruptions in the sentence formulation of school-age children

with specific language impairment. International Journal of Language and Communication Disorders.

Friedmann, N., & Novogrodsky, R. (2007). Is the movementdeficit in syntactic SLI related to traces or to thematic roletransfer? Brain and Language, 101, 5063.

Gathercole, S., & Baddeley, A. (1990). Phonological memorydeficits in language disordered children: Is there a causalconnection? Journal of Memory and Language, 29, 336360.

Gerken, L. A., & McGregor, K. (1998). An overview of prosodyand its role in normal and disordered child language. American

Journal of Speech-Language Pathology, 7, 3848.Gillam, R., & Pearson, N. (2004). Test of Narrative Language.

Austin, TX: Pro-Ed.Goffman, L. (2004). Kinematic differentiation of prosodic cate-

gories in normal and disordered language development. Journalof Speech, Language, and Hearing Research, 47, 10881102.

Guo, L.-Y., Tomblin, J. B., & Samelson, V. (2008). Speechdisruptions in the narratives of English-speaking children withspecific language impairment. Journal of Speech, Language,and Hearing Research, 51, 722738.

Hammill, D., Brown, V., Larsen, S., & Wiederholt, J. (1994).

Test of Adolescent and Adult Language, Third Edition. Austin,TX: Pro-Ed.

Hammill, D., & Newcomer, P. (2008). Test of LanguageDevelopmentIntermediate, Fourth Edition. Austin, TX:Pro-Ed.

Hedrick, D., Prather, E., & Tobin, A. (1995). Sequenced In-ventory of Communication Development, Revised Edition.Los Angeles: Western Psychological Services.

Hoffman, L., & Gillam, R. (2004). Verbal and spatial informationprocessing constraints in children with specific language im-pairment. Journal of Speech, Language, and Hearing Research,47, 114125.

Hresko, W., Reid, K., & Hammill, D. (1999). Test of EarlyLanguage Development, Third Edition. Austin, TX: Pro-Ed.

Johnson, V., deVilliers, J., & Seymour, H. (2005). Agreementwithout understanding? The case of third person singular /s/.

First L anguage, 25, 317

330.Kail, R., & Leonard, L. (1986). Word-finding abilities in

language-impaired children. ASHA Monographs, 25.Korkman, M., & Hkkinen-Rihu, P. (1994). A new classification

of developmental language disorders. Brain and Language, 47,96116.

Law, J., Garrett, Z., & Nye, C. (2004). The efficacy of treatmentfor children with developmental speech and language delay/disorder: A meta-analysis. Journal of Speech, Language, and

Hearing Research, 47, 924943.Leonard, L., Camarata, S., Pawowska, M., Brown, B., &

Camarata, M. (2008). The acquisition of tense and agreementmorphemes by children with specific language impairmentduring intervention: Phase 3. Journal of Speech, Language,and Hearing Research, 51, 120125.

Leonard, L., Ellis Weismer, S., Miller, C., Francis, D., Tomblin,

J.B., & Kail, R. (2007). Speed of processing, working memory,and language impairment in children. Journal of Speech,

Language, and Hearing Research, 50, 408428.Leonard, L., Miller, C., Deevy, P., Rauf, L., Gerber, E., &

Charest, M. (2002). Production operations and the use of non-finite verbs by children with specific language impairment. Jour-nal of Speech, Language, and Hearing Research, 45, 744758.

Leonard, L., Miller, C., Grela, B., Holland, A., Gerber, E., &Petucci, M. (2000). Production operations contribute to thegrammatical morpheme limitations of children with specificlanguage impairment. Journal of Memory and Language, 43,362378.



11/102

McGregor, K., Newman, R., Reilly, R., & Capone, N. (2002).Semantic representation and naming in children with specificlanguage impairment. Journal of Speech, Language, and

Hearing Research, 45, 9981014.McNamara, M., Carter, A., McIntosh, B., & Gerken, L. (1998).

Sensitivity to grammatical morphemes in children with specificlanguage impairment. Journal of Speech, Language, andHearing Research, 41, 11471157.

Montgomery, J. (2000). Verbal working memory and sentencecomprehension in children with specific language impairment.

Journal of Speech, Language, and Hearing Research, 43,293308.

Newcomer, P., & Hammill, D. (2008). Test of LanguageDevelopmentPrimary, Fourth Edition. Austin, TX: Pro-Ed.

Rapin, I. (1996). Developmental language disorders: A clinicalupdate. Journal of Child Psychology and Psychiatry, 37,643656.

Rapin, I., & Allen, D. (1983). Developmental language disorders:Nosologic considerations. In U. Kirk (Ed.), Neuropsychology oflanguage, reading and spelling (pp. 155184). New York:Academic Press.

Rapin, I., & Allen, D. (1987). Developmental dysphasia andautism in preschool children: Characteristic and subtypes. InJ. Martin, P. Fletcher, R. Grunwell, & D. Hall (Eds.), Pro-ceedings of the First International Symposium on SpecificSpeech and Language Disorders in Children (pp. 2035).London: AFASIC.

Rescorla, L. (1989). The Language Development Survey: Ascreening tool for delayed language in toddlers. Journal of Speech and Hearing Disorders, 54, 587599.

Rescorla, L. (2005). Age 13 language and reading outcomes inlate-talking children. Journal of Speech, Language, and Hearing

Research, 48, 459472.Reynell, J., & Gruber, C. (1990). Reynell Developmental Lan-

guage Scales ( U.S. Edition). Los Angeles: Western Psycho-logical Services.

Rice, M. (2003). A unified model of specific and general language

delay: Grammatical tense as a clinical marker of unexpectedvariation. In Y. Levy & J. Schaeffer (Eds.), Language compe-tence across populations: Toward a definition of SLI(pp. 6394).Mahwah, NJ: Erlbaum.

Rice, M., Taylor, C., & Zubrick, S. (2008). Language outcomes of7-year-old children with or without a history of late languageemergence at 24 months. Journal of Speech, Language, and

Hearing Research, 51, 394407.Rice, M., & Wexler, K. (1996). Toward tense as a clinical marker

of specific language impairment in English-speaking children.Journal of Speech, Language, and Hearing Research, 39,12391257.

Rice, M., & Wexler, K. (2001). Rice/Wexler Test of Early Gram-matical Impairment. San Antonio, TX: The PsychologicalCorporation.

Rice, M., Wexler, K., & Hershberger, S. (1998). Tense over time:

The longitudinal course of tense acquisition in children withspecific language impairment. Journal of Speech, Language,and Hearing Research, 41, 14121431.

Semel, E., Wiig, E., & Secord, W. (2003). Clinical Evaluationof Language Fundamentals, Fourth Edition. San Antonio, TX:The Psychological Corporation.

Shimpi, P., Gmez, P., Huttenlocher, J., & Vasilyeva, M. (2007).Syntactic priming in 3- and 4-year-old children: Evidence forabstract representations of transitive and dative forms. Devel-opmental Psychology, 43, 13341346.

Thal, D. (2005, November). Early detection of risk for lan-guage impairment. Paper presented at the Annual Conventionof the American Speech-Language-Hearing Association,San Diego, CA.

Thal, D., Tobias, S., & Morrison, D. (1991). Language and

gesture in late talkers: A 1-year follow-up. Journal of Speechand Hearing Research, 34, 604612.Tomblin, J. B., Mainela-Arnold, E., & Zhang, X. (2007).

Procedural learning in adolescents with and without specificlanguage impairment. Language Learning and Development, 3,269293.

Tomblin, J. B., Records, N., Buckwalter, P., Zhang, X., Smith,E., & OBrien, M. (1997). The prevalence of specific languageimpairment in kindergarten children. Journal of Speech, Lan-

guage, and Hearing Research, 40, 12451260.Tomblin, J. B., & Zhang, X. (2006). The dimensionality of

language ability in school-age children. Journal of Speech,Language, and Hearing Research, 49, 11931208.

Ullman, M., & Pierpoint, E. (2005). Specific language impairmentis not specific to language: The procedural deficit hypoth-esis. Cortex, 41, 399433.

van Daal, J., Verhoeven, L., & van Balkom, H. (2004). Sub-types of severe speech and language impairments: Psychometricevidence from 4-year-old children in the Netherlands. Journal ofSpeech, Language, and Hearing Research, 47, 14111423.

van der Lely, H. (1998). SLI in children: Movement, economy, anddeficits in the computational-syntactic system. Language

Acquisition, 7, 161192.van der Lely, H., & Battell, J. (2003). Wh-movement in children

with grammatical SLI: A test of the RDDR hypothesis. Lan-guage, 79, 153181.

Wagner, L. (2001). Aspectual influences on early tense compre-hension. Journal of Child Language, 28, 661681.

Wiig, E., Secord, W., & Semel, E. (2004). Clinical Evaluationof Language FundamentalsPreschool, Second Edition.San Antonio, TX: Psychological Corporation.

Wilson, B., & Risucci, D. (1986). A model for clinical-quantitative

classification. Generation I: Application to language-disorderedpreschool children. Brain and Language, 27, 281309.

Wolfus, B., Moscovitch, M., & Kinsbourne, M. (1980). Sub-groups of developmental language impairment. Brain and

Language, 10, 152171.World Health Organization. (2005). International Statistical

Classification of Diseases and Related Health Problems, TenthRevision. Geneva, Switzerland: Author.

Zimmerman, I., Steiner, V., & Pond, R. (2002). PreschoolLanguage Scale, Fourth Edition. San Antonio, TX: ThePsychological Corporation.

Received August 27, 2008Accepted October 21, 2008DOI: 10.1044/1058-0360(2008/08-0064)

Contact author: Laurence B. Leonard, Department of Speech,Language, and Hearing Sciences, 500 Oval Drive, HeavilonHall, Purdue University, West Lafayette, IN 47907.E-mail: [email protected].



12/102


13/102

Consensus Auditory-Perceptual Evaluation

of Voice: Development of a StandardizedClinical Protocol

Gail B. KempsterRush University, Chicago

Bruce R. GerrattUniversity of California, Los Angeles

Katherine Verdolini Abbott

University of Pittsburgh, Pittsburgh, PA

Julie Barkmeier-KraemerUniversity of Arizona, Tucson

Robert E. HillmanMassachusetts General Hospital, Boston

Purpose: This article presents the developmentof the Consensus Auditory-Perceptual Evalua-tion of Voice (CAPE-V) following a consensusconference on perceptual voice quality mea-

surement sponsored by the American Speech-Language-Hearing Associations Special InterestDivision 3, Voice and Voice Disorders. TheCAPE-V protocol and recording form weredesigned to promote a standardized approach toevaluating and documenting auditory-perceptualjudgments of vocal quality.

Method: A summary of the consensus conferenceproceedings and the factors considered by theauthors in developing this instrument are included.Conclusion: The CAPE-V form and instructions,

included as appendices to this article, enableclinicians to document perceived voice qualitydeviations following a standard (i.e., consistentand specified) protocol.

Key Words: Consensus Auditory-PerceptualEvaluation of Voice, voice, voice assessment

The Consensus Auditory-Perceptual Evaluationof Voice (CAPE-V) is a clinical and research tooldeveloped to promote a standardized approach to

evaluating and documenting auditory-perceptual judgments

of voice quality. The tool was created as a direct outcomeof the Consensus Conference on Auditory-Perceptual Eval-uation of Voice, held in June 2002 and sponsored by theAmerican Speech-Language-Hearing Associations (ASHA)Special Interest Division 3, Voice and Voice Disorders andthe University of Pittsburgh. The purpose of this article is todocument the development of the CAPE-V protocol andform, and provide a rationale for each of the elements in-cluded in the protocol.

The consensus conference brought together an interna-tional group of voice scientists, experts in human perception,and speech-language pathologists to explore solutions to a

long-standing need in clinical voice pathology: to applyscientific evidence about psychophysical measurement tothe clinical practice of judging auditory-perceptual featuresof voice quality. (See Appendix A for a list of conference

participants.) Following 2 days of presentations and discus-sion, recommendations from these participants informed andguided the development of the CAPE-V tool. The CAPE-Vauthors (the authors of this article) approached the task of

psychophysical measurement and the scaling of voice qual-ity by adhering to the consensus opinions of scientists andclinicians. From its inception, the CAPE-V was intendedto become a standardized protocol, useful to clinicians andresearchers, that incorporates multiple recommendations for

best practices in assessing perceived abnormal vocal qual-ity (Barkmeier, Verdolini, & Kempster, 2002). The wordstandardized is used throughout this article to refer to a

Clinical Focus

American Journal of Speech-Language Pathology Vol. 18 124132 May 2009 A American Speech-Language-Hearing Association

1058-0360/09/1802-0124

124


14/102

procedure that is administered and scored in a consistentway; it does not here denote norm-referencing.

The continuum of normal to abnormal voice quality is in-extricably related to vocal health. While auditory-perceptual

judgments of voice and speech can never be accomplished

with perfect validity or reliability (Gerratt & Kreiman, 2000;Gerratt, Kreiman, Antonanzas-Barroso, & Berke, 1993;Kent, 1996; Kreiman & Gerratt, 1998, 2000; Kreiman, Gerratt,Kempster, Erman, & Berke, 1993; Kreiman, Gerratt, Precoda,& Berke, 1992; Shrivastav, 2006; Shrivastav, Sapienza, &

Nandur, 2005), perceptual appraisal of voice quality remainsa key standard for judgment of vocal impairments, bothfor patients who experience vocal changes and for the cli-nicians who treat them (Carding, Carlson, Epstein, Mathieson,& Shewell, 2000; Hirano, 1981; Oates & Russell, 1998;Wilson, 1987). Simply stated, auditory-perceptual measuresof voice quality define the presence or absence of a voicedisorder clinically. Voice clinicians who treat these patientsmake auditory-perceptual judgments. Thus, there is a clear

need for a way to make such judgments that is sound theo-retically, is clinically meaningful, and can be consistentlyadministered.

Consensus Conference Issues and Summary

The 2-day conference began with a statement of theproblem, that of creating valid and reliable measures ofauditory-perceptual features of voice quality. Four invitedscientists explored issues surrounding the difficult task of

psychophysical measurement and scaling, as understoodfrom relevant areas of human perception. The presentationsreviewed the historical background of auditory-perceptualevaluation of voice and speech and described the state of theart in human auditory perception, with particular emphasison how such information might affect the auditory-perceptualassessment of voice disorders. Several voice researchers,from the United States and elsewhere, added informationfrom their investigations of voice quality. Finally, a reporton routine clinical practice in the United States was includedto relate current practice patterns to conclusions drawn fromthe scientific discussion. Throughout the conference, thescientists and clinicians reacted to the clinical and researchconundrums in auditory-perceptual judgments of voice qual-ity and the challenge of developing a new assessment in-strument. At the conclusion of the conference, the authorsof this article collaborated to draft the CAPE-V form and

procedures.

Auditory-Perceptual Evaluation: Exploringthe Elusive Ideal

Raymond Kent provided a broad review of auditory-perceptual assessment of voice and speech and outlinedthe challenges and assumptions associated with establishingan ideal perceptual evaluation method. In voice assess-ment, such an ideal method would (a) provide a reliablemeans of differentiating normal and disordered voices, andtracking changes in a patients vocal status across time;(b) correlate with underlying pathophysiology and objectivemeasures; and (c) be clearly established, including type of

scale(s), type and amount of user trainingneeded, and whetherto use anchors in training. The establishment of an idealmethod also requires that some well-known obstacles beovercome. These include the lack of standard terminologyfor describing or scaling disordered voice quality, the ab-

sence of a standard definition of normal voice, inherentpoor reliability of auditory-perceptual judgments of voicequality, and inherent variability of an individuals voice

production. Although Kent conceded that the perceptualassessment of voice quality has been an uncertain endeavor,vexed by disagreements among authorities and variabilityin data, he cited reasons for some optimism, based on agrowing international interest in developing a standardized

procedure, as evidenced by the international representationat the Consensus Conference. He also pointed to emergingconsensus points on some basic issues, including what kindof scale to use and how many and which attributes to rate.Moreover, recent demonstrations suggest that computermodeling and interactive synthesis of disordered voice qual-

ity can assist in developing improved methods for auditory-perceptual assessment (Callan, Kent, Roy, & Tasko, 2000;Gerratt & Kreiman, 2001; Kreiman & Gerratt, 2000).

Psychoacoustic Principles and Human Perception

Lawrence Feth reviewed current psychoacoustics-basedperspectives on human perception and the discrimination ofsound (Houtsma, 1995; Zwicker, Fastl, & Frater, 1999). He

presented a brief review of the anatomy and physiology ofthe ear, as well as an overview of how sound is processedgoing from peripheral to central auditory mechanisms. Hedescribed (a) the peripheral influences of the outer andmiddle ear as manifested by the audibility curve, (b) thefrequency selectivity/critical bandwidth processing and in-

tensity compression that are initially the result of cochlearmorphology and biomechanics, and (c) integration of acousticinformation (e.g., spectral integration) that takes place athigher levels of the central nervous system. Feth also sum-marized what is currently known about how humans perceiveand discriminate acoustic parameters of sound. Most of thework in this area has focused on pitch and loudness per-ception, which has influenced the scaling methods and theo-retical constructs for both of these phenomena. Much lesseffort has been expended in formally studying the perceptionof sound quality, primarily because it is a more complex(multidimensional) and difficult to quantify perceptual phe-nomenon. Two sounds that are judged to have equal pitchand loudness but can still be discriminated from each other

are said to differ in timbre orquality. By way of example,Feth briefly summarized some of the work by Zwicker et al.(1999) in which they attempted to explore the perception ofquality-related concepts such as sharpness, pleasantness,

fluctuation, strength, and roughness.

Psychophysical Issues Related to Scaling

George Gescheider and Lawrence Marks each gavepresentations dealing with the psychophysical bases ofperceptual scaling and measurement (Gescheider & Marks,2002; Marks & Algom, 1998). Gescheider briefly reviewed

Kempster et al.: Consensus Auditory-Perceptual Evaluation of Voice 125


15/102

the classic work of some well-known pioneers in psycho-physics (Fechner, Weber, Stevens, and Thurstone) in dis-cussing basic approaches for determining absolute anddifference (just noticeable differences or difference limens)thresholds for sensory systems. He concluded with an over-

view of additional scaling methods, including partition,ratio, and multidimensional approaches.

Marks addressed the methodological issues in perceptualscaling and measurement more specifically, explaining thatthe general process of psychophysical analysis involves(a) definition of the stimuli (What properties of stimuliare pertinent to perception?), (b) definition of the perceptualexperiences (What are the attributes/features/dimensionsof perception and how do these attributes/features/dimensionsinterrelate?), and (c) determination/modeling of the pro-cesses that relate percepts to stimuli (sensory, decisional,cognitive processes: encoding, transforming, recoding,etc.).

Marks also commented on the concepts of internal and

external validity, sensitivity, and reliability of scaling pro-cedures, as well as differences between direct as comparedwith indirect scaling methods. Direct methods, which areconsidered more appropriate than indirect approaches forclinical applications, make use of interval or ratio scaling

procedures. To optimize direct scaling, consideration shouldbe given to the actual range and distribution of the stimulibeing used, whether standard or anchor stimuli are employed,whether there are sequential/order effects, and whethertraining improves performance (Gescheider & Marks, 2002;Marks & Algom, 1998).

Marks made two specific recommendations with respectto developing a clinical instrument for auditory-perceptualassessment of voice quality. First, he recommended usingnumerical rating scales with at least 15 subcategories/divisions or, alternatively, employing continuous graphical/visual analog scales. Second, he recommended that the lo-cation of anchors (e.g., normal or most severe) be adjustedto provide extra room at the ends of the scale to avoid endeffects.

Current Practice

Reports from international experts present at the confer-ence reviewed the utilization of auditory-perceptual scales inclinical practice and research, including interactive trainingmodels and the use of training scales and anchors (Chan &Yiu, 2002; Oates & Russell, 1998) and other formalized

perceptual scaling instruments and procedures including

Vocal Profile Analysis by Laver (Carding et al., 2000) andthe Stockholm Voice Evaluation Approach (Hammarberg,2000). Carding reviewed current methods in Britain andnoted that while most clinicians in the United Kingdomare trained in Lavers Vocal Profile Analysis, the GRBAS(grade, rough, breathy, asthenic, and strained) method(Hirano, 1981) is recommended as the minimum standard for

practicing voice clinicians in the United Kingdom. Severalparticipants referred to other influential sources of informa-tion related to the perception of vocal quality: Kreiman et al.(1993) and DeBodt, Wuyts, Van de Heyning, and Croux(1997). Kreiman et al. identified 57 different perceptual

schemes for voice assessment and concluded that the mostwidely used was the Buffalo Voice Profile (Wilson, 1987).Work in Belgium by DeBodt and his colleagues includesclinical recommendations about appropriate use of variousoptions based on a review of contemporary perceptual rating

scales.

Consensus Points

The conference attendees agreed that there is no single,best way to approach the task of measuring perceived vocalquality. The current knowledge base is inadequate for de-signing a clinical tool that resolves all of the relevant scien-tific issues. Indeed, efforts to do so have reflected an arrayof problems of reliability, utility, and validity, and theselimitations are also true in the development of the CAPE-V.

Nonetheless, the CAPE-V authors incorporated multipleperspectives, from scientific data to clinical practice, todevelop both a protocol to follow and a form to document

auditory-perceptual features of abnormal quality. Confer-ence participants agreed that constructing a consistent andspecified set of evaluation procedures and a documentationformat would, at a minimum, improve communication andconsistency among clinicians.

In this context, the authors agreed on the followingorienting principles:

1. Perceptual dimensions should reflect a minimal set ofclinically meaningful, perceptual voice parameters.

2. Procedures and results should be obtainable expediently.

3. Procedures and results should be applicable to a broadrange of vocal pathologies and clinical settings.

4. Ratings should be demonstrated to optimize reliability

within and across clinicians through later validationstudies.

5. Ultimately, exemplars may be considered for future useas anchors and possibly for training.

Specific Elements of the CAPE-V

The CAPE-V instructions are included as Appendix B,and the form for documenting the assessment is presented inAppendix C.

Tasks

The CAPE-V stipulates that the individual whose voice is

to be assessed (hereafter referred to as thepatient

) perform

three specific vocal tasks: (a) sustain the vowels /a/ and /i/three times each; (b) read six specific sentences with dif-ferent phonetic contexts; and (c) converse naturally in re-sponse to the standard question (Tell me about your voice

problem).Rationale for the tasks. The first task elicits vowel pro-

longations. Vowel prolongations (at a steady and comfortablepitch level) provide an opportunity to listen to a patientsvoice without articulatory influences. Vowels can also beanalyzed acoustically, for which some normative data areavailable. The second task elicits six sentences of varied



16/102

speech contexts from which to assess different elementsof vocal quality. Sentence 1 (The blue spot is on the keyagain) is a commonly used stimulus sentence to exam-ine the coarticulatory influence of three vowels (/a, i, u/ ).Sentence 2 (How hard did he hit him?) provides a con-

text to assess soft glottal attacks and voiceless to voicedtransitions. Sentence 3 (We were away a year ago) fea-tures all voiced phonemes and provides a context to judge

possible voiced stoppages/spasms and ones ability tolink (i.e., maintain voicing) from one word to another.Sentence 4 (We eat eggs every Easter) includes severalvowel-initiated words that may provoke hard glottal attacksand provides the opportunity to assess whether these oc-cur. Sentence 5 (My mama makes lemon jam) includesnumerous nasal consonants, thus providing an opportunityto assess hyponasality and possible stimulability for res-onant voice therapy. Finally, Sentence 6 (Peter will keepat the peak) contains no nasal consonants and providesa useful context for assessing intraoral pressure and pos-

sible hypernasality or nasal air emission. The third taskelicits conversational speech and is the most important andrelevant to both patient and clinician. Although, in theCAPE-V protocol, conversation is assessed after the vow-els and sentences, it is expected that this aspect of the

patients voice is under close observation throughout theevaluation session.

Quality Features to be Assessed

The CAPE-V protocol specifies six quality features tobe evaluated consistently and allows flexibility to add otherperceptual features of interest. The six voice quality featuresselected for consistent appraisal are labeled and defined asfollows:

Overall Severity: global, integrated impression of voicedeviance

Roughness: perceived irregularity in the voicing source

Breathiness: audible air escape in the voice

Strain: perception of excessive vocal effort (hyperfunction)

Pitch: perceptual correlate of fundamental frequency

Loudness: perceptual correlate of sound intensity

Rationale for the quality features. Despite much debateover the description, validity, and independence of any listof voice quality features, these six have consistently ap-

peared in both national and international voice literature

for decades (DeBodt et al., 1997; Fairbanks, 1960; Hirano,1981; Wilson, 1987). Thus, the rationale for including thesesix voice quality features is the belief that both cliniciansand researchers find these attributes meaningful. Anothercommon descriptor, hoarse, was excluded from the list ofterms because the authors agreed with Fairbanks (1960)thathoarseness is perceived by many as a combinationofroughness and breathiness. The CAPE-V form alsoincludes two unlabeled scales. These allow the clinician todocument other salient perceptual features of a patientsvoice, such as degree of nasality, spasm, tremor, intermittentaphonia, falsetto, glottal fry, or weakness.

Scale

A 100-mm line scale with unlabeled anchors, commonlyknown as a visual analog scale, is used to assess each ofthe six quality features. The leftmost portion of the scale

reflects normal voice (in the case of judging severity, pitch,or loudness) or none of the quality being judged (in the caseof roughness, breathiness, and strain). The right end ofthe scale is to reflect the listeners judgment of the mostextreme example of deviance. A tick mark for each of thethree tasks, with the subscript 1 (for vowels), 2 (for sen-tences), and/or 3 (for conversation) is drawn onto the scaleto reflect a listeners judgment for each scale. Measurementfrom the left end of the scale to each tick mark, in milli-meters, is denoted on the blank to the far right of the scale(___/100).

Rationale for the scale. Marks recommends that auditory-perceptual judgments of voice quality be made on a visualanalog scale (or set of scales), using open-ended anchor

points at either end as a way to inhibit end effects of thescale. Visual analog scales are easy for raters to use andappear to have become more commonplace in voice researchin the past 2 decades.

Verbal Descriptor Degree of Deviance

While the primary measurement index is an interval scaleprovided by the 100-mm visual analog line, the CAPE-V alsoincludes the ordinal ratings of mild, moderate, and severe,printed below the measurement line, to serve as a supple-mental severity indicator. These qualitative terms are po-sitioned in a nonequidistant fashion, based on Markssrecommendations, and reflect the range of voice severityusing terminology more familiar to clinicians than thediscrete intervals measured on the 100-mm visual analogscale.

Additional CAPE-V Elements

A nominal rating judgment allows the clinician to clas-sify the consistency or intermittent presence of the voicequality feature within and across evaluation tasks. Sectionsdevoted to resonance or other features supplement theCAPE-V protocol by allowing other salient descriptors todocument a patients voice quality. This flexibility is neededto capture the spectrum of voice disorders and associatedconditions or features. The list of terms provided on the formis not inclusive, meant only as examples of specific featuresthat may help describe auditory-perceptual attributes.

Rating Procedures

The CAPE-V judgments are intended to reflect theclinicians direct observations of the patients performanceduring the evaluation and should not take into account

patient report or other sources. Standard audio-recordingprocedures should be used, such as recording in a quietenvironment and using a standard mouth-to-microphonedistance with the highest possible sampling rate for digitalconversion. If a patient returns following an initial assess-ment, the clinician may compare the initial voice sample and

Kempster et al.: Consensus Auditory-Perceptual Evaluation of Voice 127


17/102

CAPE-V ratings directly to any subsequent recordings, tooptimize the internal consistency or reliability of repeatedsequential ratings, particularly for assessing treatment out-comes. As always, clinicians are encouraged to minimize

bias in all ratings.

The CAPE-V form and instructions have been availableto affiliates of Special Interest Division 3 on the protected

portion of the divisions Web site since 2003. The tool waspresented at national conventions as early as 2002 (Barkmeier,2003; Barkmeier et al., 2002; Hillman, 2003; Shrivastav,Kempster, & Zraick, 2006; Zraick et al., 2007). The instru-ment is already used in more than 20 clinics and somelaboratories throughout the United States, and using theCAPE-V protocol as directed has been shown to add nomore than a few minutes to a voice evaluation session(M. Spencer, personal communication, June 16, 2005). Anational, multi-institutional validation study examining thereliability of the instrument has also begun (Zraick et al., 2007).

Concurrent Validity and the CAPE-V

A masters thesis (Berg & Eden, 2003) directed byHammarberg and Holmberg compared aspects of theCAPE-V to the Stockholm Voice Evaluation Approach on

patients with three different voice pathologies (E. Holmberg,personal communication, December 1, 2003). This studyinvolved a translation of the CAPE-V into Swedish. Theauthors determined that intra- and interrater reliability wasacceptably high in both protocols, and no obvious differ-ences were found between the two approaches in terms oflistener variability. Both protocols were able to separatethe three disorders from each other and showed significant

pre-to-posttreatment changes in voice quality.Karnell et al. (2007) published a preliminary report com-

paring the reliability of clinician-based auditory-perceptualjudgments using the CAPE- V to those made with theGRBAS voice-rating scheme (Hirano, 1981) and two otherquality of life scales. Among other findings, Karnell et al.found comparable estimates of interrater reliability forthe two scales, both at high levels. They suggest that theCAPE-V may offermore sensitivity to small differenceswithin and among patients than the GRBAS scale (p. 1).

A second preliminary investigation has suggested thatthe CAPE-V results meet or exceed the GRBAS in measure-ment reliability (Zraick et al., 2007).

The CAPE-Vs similarity to the GRBAS scale is obviousto anyone familiar with both scales. In fact, the CAPE-V usesall of the GRBAS percepts (except aesthenic) for judging

voice quality, and the definitions of the quality features arealso similar. However, three important factors discriminatethe CAPE-V from the GRBAS scale. First, the GRBAS hasno published, standardized protocol to follow in English.The Hirano (1981) reference most often cited for the GRBAS

provides no guidelines for clinical administration, speechmaterial, or rating calibration. In contrast, the CAPE-Vincludes a specific protocol that designates the tasks, pro-cedures, and scaling routine, toward the larger goal ofimproving the consistency of clinical assessment from oneclinician to another, without excessive demands on cliniciantime or learning. Second, the CAPE-V provides interval

scale measures of voice quality by incorporating millimetermeasures on vis

aslha may 1 2009, volume 18, issue 2

Documents