BEYOND 14 GRAMMATICAL MORPHEMES: WHAT DO SPEECH-LANGUAGE PATHOLOGISTS NEED TO CONSIDER IN THE EVALUATION OF GRAMMATICAL SKILLS?
In the 1960s, Roger Brown, a developmental psychologist, and his colleagues set about to characterize the grammatical development of three typically developing children to whom they referred as Adam, Eve, and Sarah. Adam, Eve, and Sarah were from a diverse set of families- having an African American father who was a preacher, a White father who was a graduate student, and a White working-class father, respectively. The research team visited the children in their homes quite frequently (biweekly, monthly), audio-recorded the children's talking, and returned to the laboratory to orthographically transcribe the children's talk. Today, the ease with which clinicians and researchers can audio/video record children and use computers for transcription and analysis can make us quickly forget what a challenge it was to do these things several decades ago!
More than 40 years later, what is most enduring from Brown's work, particularly in language development textbooks, are 14 grammatical morphemes and mean length of utterance (MLU; Brown, 1973). Brown's research team documented children's increasing use of 14 morphemes by reporting the percent inclusion of each grammatical morpheme in obligatory contexts. They also captured changes in the length of children's utterances, using a metric they devised and called mean length of utterance in morphemes (see also Eisenberg, Fersko, & Lundgren, 2001). There are fortunate upsides to Brown's work-for example, metrics to capture growth in grammatical skills and an initial picture of grammatical development. Brown's influence on research in typical as well as atypical language learning is readily visible and extensive. His research team's efforts spawned great interest in children's language development across multiple disciplines, including communication sciences and disorders, developmental psychology, and linguistics. Forty years of research on typical and atypical language learning has provided a well-illustrated (albeit incomplete) picture of grammatical development. Although researchers continue to analyze MLU and use of morphemes in obligatory contexts, they have developed additional metrics that are used in research and in clinical practice to capture grammatical development (e.g., Scarborough, 1990).
In the clinical practice of speech-language pathology, there has been an unfortunate downside to Brown's work-not at all attributable to Brown and his colleagues but to how their work has been interpreted-that is, a narrow view of grammatical development. For too many clinicians, grammatical development begins and ends with (or does not move far beyond) MLU and 14 grammatical morphemes. What is the impact of this narrow interpretation of grammatical development on clinical practice? Clinicians are unable to address adequately the language needs of children whose deficits in the grammatical aspect of language are central to their language- learning difficulties. Only by stepping back and taking a much broader view of grammatical development will Brown's influence on language development have the impact on clinical practice that it should.
The motivation for Brown's investigation of the language development of Adam, Eve, and Sarah was not clinical. Rather, the motivation most simply was to understand the phenomenon of typical language development. But because clinical practice in speech-language pathology draws on the basic research in many other disciplines, clinical practice has been influenced extensively by Brown's work. The purpose of this introductory article and the remaining articles in this issue of Topics in Language Disorders is to challenge clinicians to think broadly about grammatical development so as to address the full range of grammatical deficits that have an adverse impact on the overall communicative proficiency and academic and vocational achievements of children (and adults) with primary or secondary language impairment.
WHY THOSE 14 GRAMMATICAL MORPHEMES?
A careful reading of Brown's book (1973), A First Language, reveals some critical aspects of his research that are overlooked too often and that do not make their way into textbooks. In characterizing grammatical development, any number of grammatical structures could have been selected for study. A cursory glance at the list of inflectional and derivational morphemes Ingram (1991) provided (Table 1) illustrates the pool of possibilities that might have been studied. The grammatical morphemes Brown chose to study were those that could be studied reliably. Research data must include sufficient exemplars for interpretation, as well as agreement across researchers. Agreement and frequency made these 14 morphemes suitable for investigation. First, the research team was in agreement as to when these 14 morphemes should have been used by the children (i.e., agreement on obligatory contexts), and they were able to establish reliability in their data coding. Second, the language sample data collected from Adam, Eve, and Sarah provided sufficient contexts in which these morphemes were obligatory. Thus, it was possible to characterize percent correct in obligatory contexts. The research team could formulate a valid picture of developmental performance and change across the preschool years for these 14 grammatical morphemes (Table 1).
What is critical for clinicians to keep in mind is that there was nothing special that was identified a priori about the morphemes that Brown chose to investigate. For example, Brown did not argue that these were the most important grammatical morphemes for children to master, nor that these morphemes presented the greater challenge to typical (or atypical) language learners. Rather, these 14 were the morphemes for which the child language transcripts provided sufficient and reliable data on which a story of development could be built. These morphemes (and MLU) were a great starting point to the characterization of grammatical development, but, importantly, just a starting point. Similarly, MLU and 14 grammatical morphemes may provide a great starting point for clinicians. A problem ensues when this starting point becomes the ending point as well.
These 14 morphemes were identified as markers of growth in typical language learners. They are neither the sum total of grammatical morphology nor the curriculum for grammatical intervention. Mastery of these morphemes does not provide evidence that a child has "mastered" the grammatical domain of language. Importantly, if clinicians are to address adequately the grammatical limitations of children with primary or secondary language impairment, a much broader perspective on grammatical development must be adopted.
WHAT IS A BROAD PERSPECTIVE ON GRAMMATICAL DEVELOPMENT?
A cursory glance at the more than 1,500 pages used by Quirk, Greenbaum, Leech, and Svartik (1985) to detail a grammar of English reveals that the grammatical proficiency of the typical human mind encompasses an extensive body of knowledge, albeit knowledge that is largely inaccessible to conscious awareness or manipulation. Grammatical proficiency, or primary language knowledge, entails the ability to seamlessly and fluently generate grammatical utterances to express one's thoughts and ideas. In contrast, metalinguistic knowledge involves the ability to think about language structure as an object of thought-an ability that is not necessary for grammatical proficiency and, for most humans, is neither seamless nor fluent (Hakes, 1980). A broad perspective on grammatical acquisition requires thinking about what it is that the preschool human brain acquires that allows the child to convey his or her thoughts in grammatical utterances that vary widely in structure and that encompass an expansive lexicon.
A broader perspective on grammatical development can begin with Brown's 14 grammatical morphemes. Within this group of morphemes, there is varying information on development provided by individual morphemes. For example, the aspect marker -ing appears early in grammatical development and children quickly move toward using -ing in all present progressive contexts. In contrast third person -s (jumps, plays, chases) and regular past tense -ed (jumped, played, planted) show much more protracted developmental trajectories for typical as well as atypical language learners (Rice, Wexler, & Hershberger, 1998). The metric of percent correct in obligatory context can be informative with other grammatical morphemes, for example, nominative pronouns, such as she is my friend (Rispoli, 1998; Schuele, Haskill, & Rispoli, 2005), infinitival TO, such as the boy wanted to fly the kite (Barako Arndt & Schuele, 2012), and subject relative markers, such as the pig that fell in the mud (Schuele & Tolbert, 2001). When a grammatical structure has an obligatory context that can be identified with reasonable reliability, percent correct in obligatory contexts is the metric of choice to quantify development (as opposed to, for example, frequency of correct uses within a language sample).
But percent correct in obligatory contexts tells only part of the proficiency story. Radford (1990) argued for a broader view of proficiency. Proficiency presumes the productive use of a morpheme-with a variety of bare stems or in a variety of utterance contexts. In addition, proficiency presumes that a morpheme appears only in the contexts in which it is expected (Table 2). It is possible for a child to use a morpheme consistently in obligatory contexts (e.g., 90% accuracy) yet not show evidence of these other measures of proficiency.
The metric of percent correct is not useful for those grammatical structures for which there is no obligatory context. A case in point is complex syntax, the production of dependent clauses (see Barako Arndt & Schuele, 2013). Obligatory context does not work in this instance because there are no obligatory contexts for infinitival complements, subject relatives, and so on. A child can choose to use a particular complex syntax structure or not. So, if a child does not produce a particular complex syntax structure in a particular language sample, one cannot conclude that the child cannot produce the structure, only that the child did not. Thus, to capture grammatical performance and growth over time, clinicians need to consider the nature of the grammatical structure of interest and select a metric accordingly.
Taking a broader perspective of grammatical development requires thinking not just about inflectional morphemes (a typical area of inquiry) but derivational morphology (a less typical area of inquiry) as well. Literacy research over the last decade has underscored the importance of derivational morphological knowledge to reading comprehension (Green, 2009; also see Volume 33, Issue 1, of Topics in Language Disorders, edited by Smith, 2013). Speech-language pathologists (SLPs) tend to be much more focused on grammatical morphology learning than derivational morphology learning, but both are essential for a productive grammatical system that allows a child to communicate effectively in spoken as well as written (writing, reading) language. Many clinicians have tuned their ears to hear "goed" and "rided" as evidence that a child is figuring out marking past tense, but fewer clinicians have tuned their ears to productions such as I drew a hotdogular person (a person in the shape of a hot dog; as in circular, rectangular, Daniel, 4-year-old) as evidence that a child is figuring out how to turn a noun into an adjective (see Clark, 2009 discussion on legal and illegal coinages).
In sum, a broader view of grammatical development requires a clinician to think about all that a child needs to know to be grammatically proficient. The grammatically proficient child can produce a wide array of grammatical structures. And, the grammatically proficient child can use those grammatical structures with an amazingly large lexicon. Sources such as Scarborough's (1990) "Index of Productive Syntax" provides language specialists with an accessible framework for thinking about the scope of grammatical knowledge that a child acquires.
COMING TO THE GRAMMATICAL TABLE: WHAT DOES THE SLP BRING?
If competent clinical practice requires a clinician to think broadly about grammatical development, a clinician must have thorough knowledge of the grammatical structure of language (i.e., English or whatever the language of assessment and intervention is) and the ability to think about the structure of language, that is, metalinguistic ability. Without this knowledge and ability, a clinician's efforts to address the grammatical assessment and intervention needs of individual children with language impairments will fall short. Several years ago, Long (1996) reflected upon the insufficient grammatical knowledge of SLPs in training, and the situation does not appear radically different today.
Faculty in communication sciences and disorders who teach language development and language disorders courses collectively bemoan the lack of grammatical knowledge of far too many undergraduate and graduate students in speech-language pathology. It is likely that a substantial shift in preprofessional preparation is needed to address these knowledge deficits. Undergraduate curricula in communication sciences and disorders long have included a course in speech science, as well as a course in hearing science. Yet, a comparable course in language science has not been routine. Most assuredly if SLPs need a thorough knowledge of grammar, then coursework decisions must reflect that expectation. For practicing clinicians, gaining expertise in grammar may be achievable only with individual study. Professional development workshops typically are not focused on this topic.
For now, interested readers might begin with Long (1996) and use the appendix to self-assess their current knowledge base. Having an idea of what one knows and does not know, the next step involves identification of learning resources. By far the most extensive source of information on English grammar is Quirk et al. (1985), A Comprehensive Grammar of the English Language. However, reading this book, even excerpts, is likely palatable only for grammar aficionados.
Fortunately, several more palatable resources are available for those wishing to increase their grammatical knowledge and metalinguistic skills. Quirk and Greenbaum (1973) provides a condensed version of the content in Quirk et al. (1985), which is quite readable. Moats (2003, 2010) provides resources primarily with preprofessional teachers in mind, but the content is equally of interest to SLPs. In addition to information about spoken language, Moats' sources will boost clinicians' understanding of the orthographic structure of written language (e.g., Why is it bubble and not buble?). Parker and Riley's (2009) text, Linguistics for Non-linguists: A Primer with Exercises, provides a thorough walk through the structure of language and does not assume a priori knowledge of linguistics or language science. Justice and Ezell's (2002) text, written for SLPs, provides a good overview but is not extremely detailed. Barako Arndt and Schuele (2013) provide readers with Supplemental Digital Content (available at: http://links.lww.com/TLD/A17) that can support practice identifying child utterances with complex syntax and categorization of complex syntax tokens into types. To be sure, there are many other books and resources on this topic; the few referenced here may prove helpful as a place to start.
THIS ISSUE
The authors of the articles in this issue of Topics in Language Disorders were invited to challenge readers to broaden their perspective on grammatical development and intervention. The scholarly work of these authors clearly indicates their biases-that grammatical proficiency is critical to language and communication proficiency and that assessment and intervention must encompass a broad perspective on grammatical development.
Barako Arndt and Schuele (2013) challenge readers to think about complex syntax, particularly beginning in the preschool years. Language development textbooks are only beginning to devote a substantial amount of space to complex syntax production. Yet, the ability of a child with language impairment to gain proficiency in using a variety of complex syntax structures may be critical to academic success-from engaging in higher level verbal discourse in kindergarten to writing an analysis of the Civil War in high school. Working from a grammatical perspective, Barako Arndt and Schuele describe a framework for analyzing the complex syntax produced by preschool- and early elementary school-age children. It is a framework that emphasizes the grammatical knowledge children must gain to produce a variety of complex syntax types.
Oetting, Lee, and Porter (2013) challenge readers to consider the grammatical development of children with language impairments who are speakers of nonmainstream English dialects. Given that there are substantial differences between the grammatical structure of some nonmainstream English dialects and mainstream English dialects, some clinicians may be tempted to ignore grammatical development in nonmainstream dialect-speaking children. Oetting et al. (2013) give us reason to avoid that temptation.
Weiler (2013) challenges readers to think carefully about the features of intervention stimuli when targeting grammatical proficiency. Although Weiler focuses on factors that clinicians should think about when planning intervention to improve children's production of past tense, the implications of his article are relevant to all grammatical intervention targets.
Finally, Eisenberg (2013) challenges readers to think about the many ways clinicians can build grammatical proficiency in preschool- and early school-age children. Importantly, her perspective focuses on the production of grammatical structures in meaningful and rich communicative contexts. Typical language learners acquire grammatical structure in the context of meaningful communicative interchanges. Thus, Eisenberg argues that the same is important for children with language impairments.
As issue editor, I hope that Topics in Language Disorders readers will find that the articles in this issue help them think beyond Brown's (1973) 14 morphemes. Those morphemes are not unimportant, but there is so much more richness to consider.
-C. Melanie Schuele
Issue Editor
Vanderbilt University
REFERENCES