RESEARCHERS in the development of literacy have uncovered discernable developmental changes in reading and writing over the entire course of formal schooling from kindergarten through secondary school and beyond to the college years. To characterize the essence of change in reading, Jean Chall (1983) identified six stages: the first three have been popularized as “learning to read” and include basic decoding and building of automaticity and accuracy. Then, between the third and fourth grades, children enter the “reading to learn” period that, over time, becomes increasingly nuanced in terms of the ability to draw inference, synthesize information, and think critically. For the last quarter of the 20th century and continuing to the present, the clear emphasis in education and policy circles, and the attendant allocation of federal funds, has been on learning to read. The common belief seems to be that if basic literacy is well taught and well learned in the first 3 or 4 years of schooling, further literacy growth will occur automatically without explicit instruction—a view dubbed “the vaccination model” by Shanahan and Shanahan (2008, p. 43). The idea is that a strong dose of early instruction in the fundamentals of reading will provide protection against later literacy “disease.”
Recent national reading and writing assessment data from the 4th through 12th grades (Grigg, Donahue, & Dion, 2007; Perle & Moran, 2005; Persky, Daane, & Jin, 2003), as well as comparisons of U.S. middle- and high-school students with their peers in other developed nations, tell a different story—one that has grabbed the attention of foundations (e.g., the Carnegie Corporation of New York, the Bill and Melinda Gates Foundation) and nonprofit organizations (e.g., the Alliance for Excellent Education) concerned with the viability and end product of public education. These organizations have sponsored several analyses of why proficient levels of reading and writing elude well over half of all older students in this country (reports are accessible on the organizations’ websites) along with thoughtful papers that address what can be done (for reading, see Biancarosa & Snow, 2006; for writing, see Graham & Perin, 2007).
A common thread emerging from discussions of literacy travails in older students is that the processing of informational (i.e., expository) text1 poses a special set of problems that can only be addressed by expanding explicit literacy instruction into the higher grades. This expanded literacy instruction can be further categorized. One type of instruction would target generic reading comprehension strategies and vocabulary work associated with common academic word meanings. This level of literacy is termed intermediate literacy by Shanahan and Shanahan (2008). However, an increasing number of experts and researchers write that this approach will still fall short because it fails to help students deconstruct the content-specific linguistic features unique to curricular subjects, namely science, math, history, and literature. They argue that what is needed is disciplinary literacy—instruction that would help students appreciate “differences in how the disciplines create, disseminate, and evaluate knowledge” .... and how “these differences are instantiated in their use of language” (Shanahan & Shanahan, p. 49, emphasis added). In the literature on academic literacy, general recommendations for sensitivity to these unique features are more common than detailed descriptions of these features, and further, descriptions are more common than discussions of how this information would translate into instructional practice (with a few notable exceptions that are summarized in the last section).
This brings us to the first topic for this article, a description of three grammatical strategies used to encode information found in content texts in late elementary and secondary schools—complex nominal groups, clausal subordination, and theme and information structuring patterns. Implied in this discussion is how these features would pose comprehension challenges for any student. We then turn to the population of students with a history of language impairment and/or learning disabilities and review evidence that confirms the challenges these sentences pose for these individuals. Finally, we consider how this information could be applied in both assessment and instruction–intervention protocols when working with older students who struggle with academic language.
THE GRAMMAR OF INFORMATIONAL TEXT
Descriptions of informational language abound in linguistic and educational literature. A heuristic often used is to compare informational text with other genres—most often narrative text and oral conversational–interactional language. Moreover, because informational text is often written text, one finds frequent reference to the structural features of writing compared with speaking (a modality comparison). Both genre and modality comparisons have been summarized by Scott (1994), drawing on the work of Halliday (1987), Halliday and Hasan (1976)), and Biber (1986, 1988), among others. A recent treatment of the language of schooling in the tradition of functional linguistics (Schleppegrell, 2004) stresses three characteristics that distinguish academic texts: field (ideation: what is talked about), tenor (interpersonal stance: the relationship between speaker/hearer or writer/reader), and mode (textual: expectations for how particular text types should be organized). The characteristic of field in school texts is most often coded in lexical choices that are technical and abstract, followed by specific meanings expanded within long nominal groups. These texts also have a high lexical density (a higher proportion of nouns, verbs, and adjectives) (Fang, 2004, 2008; Schleppegrell, 2001). The characteristic of tenor is one of authority and distance, coded with such structures as declarative sentences (fewer interrogatives and imperatives), modal verbs (e.g., cannot), passive voice constructions that remove identifiable agents, and contrastive logical arguments realized with adverbial clauses (e.g., although clause 1, clause 2) (Schleppegrell, 2004). One of the major ways that the characteristic of mode (textual organization) is communicated in academic texts is through clausal subordination strategies. Lacking intonation, which serves an organizational function during an oral interaction, written school texts must rely on logical organization mechanisms. This organization is more hierarchical than in oral discourse, with extensive embedding and subordination. Another major feature of mode is the way that themes are developed. Themes appear at the beginning of clauses/sentences and then are elaborated upon at the end of the clause or sentence. Examples of selected grammatical structures used to code ideation, tenor, and mode are described in greater detail below.
From the perspective of the functional linguist, grammar is the tool for achieving the social/situation purposes of informational text. The tool offers choices in the way it can be used (there is more than one grammatical way to achieve a particular function), and the effectiveness of the informational speaker/writer depends on controlling an array of such choices. In this section, we highlight three grammatical strategies that contribute to field, tenor, and mode in academic texts: nominal groups, clausal subordination, and theme and information structuring patterns. Although our emphasis is on syntax, and the “domain” of syntax is the sentence, we recognize that additional intersentential connectives (e.g., however, otherwise) and cohesion devices (Halliday & Hasan, 1976) make important contributions to academic texts. We are also aware of interdependencies between lexical and syntactic levels of language, termed “lexical-syntax interface” by Berman and Nir (2010), but we do not address those in this article.
A good place to start is to think about the function of academic texts and how they differ from other types of texts. Subjects and themes in conversations and in narratives are most often people, with the result that subjects are named individuals and pronouns (I..., Emily..., in the new movie Avatar Jake ...). Subjects and themes in textbooks, on the other hand, are whole events/periods of history (World War II, the Wall Street crash of 1929) and processes (natural selection, echolocation, shearing). To accomplish this, terms denoting a measure (e.g., liter, byte), category (e.g., mammal, vertebrate), or social/official station (e.g., czar, president) are frequent. The raison d’être of academic texts read by students is to teach content; the major way this occurs is by explaining, describing, and comparing (Bailey, Butler, Stevens, & Lord, 2007). As a result, subject and object noun phrases (NPs) are long and complex, as this teaching function is encoded in extensive premodification and/or postmodification of the head noun of the NP. A noun can be expanded and modified almost indefinitely, allowing an author/teacher to “pack” a great deal of information into a sentence. A student reading the following sentence must somehow parse the 16-word subject NP (shown in italics) as a grammatical unit:
- The oldest known fossil skeleton of a human ancestor—a female Ardipithecus Ramidus specimen nicknamed “Ardi” (pictured)—has been found, scientists revealed yesterday. (National Geographic Kids, 2009)
In Example 1, the author is delimiting the NP head noun skeleton by teaching that a skeleton, if old enough, can be classified as a fossil; that it belonged to a female human ancestor; that fossils are classified with fancy Latin names; that it is a rare find—hence its status as a specimen; and that scientists have a playful side in nicknaming their finds. And, all this teaching takes place within the grammatical confines of the subject NP of the sentence. The head noun undergoes both premodification (the oldest known fossil) and postmodification in the form of a prepositional phrase (of a human ancestor) and an appositive construction that defines ancestor (a female Ardipithecus Ramidus specimen nicknamed “Ardi”). The postmodification places 12 words between the clausal subject (the head noun skeleton) and its predicate (has been found); the resulting interruption of the subject and verb undoubtedly contributes additional processing load for the reader. Postmodification in long NPs frequently takes the form of embedded relative clauses (reduced here to nicknamed Ardi), which have been studied extensively by psycholinguists as syntactic structures that increase processing complexity for both children and adults (e.g., Traxler, Morris, & Seely, 2002). In Example 1, a further processing challenge is the passive voice status of the first clause of the sentence. Who found the specimen is not directly stated, but readers could assume, it was the scientists identified in the second clause that revealed the findings. Passive voice is one way to adapt an impersonal and hence authoritative stance within an academic register, a feature of tenor, as discussed above.
Another prominent characteristic of nominal groups in academic texts is the process of nominalization—the process that transforms verbal information into nominal information. Example 2 below is the sentence that follows (1) above in the source text. The verb phrase has been found in (1) becomes the find in (2) and doing so contributes to the cohesion and organization (i.e., mode) of the text by summarizing information in the previous sentence and coding it as a given in the next sentence.
- The find reveals that our ancestors underwent a previously unknown stage of evolution more than a million years before Lucy, the early human ancestor specimen that walked the Earth 3.2 million years ago.
Find as a subject noun could confuse students because of its rarity in that grammatical role and complicate parsing an already complex sentence that contains three clauses and two complex NPs (a previously unknown stage of evolution; Lucy, the early human ancestor specimen that walked the Earth 3.2 million years ago).
Nominalization is prominent in both science and history texts, but with a nuanced difference. In science texts, nominalization contributes to the technical, abstract quality of the text. In the sentence “salt goes through a process of dissolution,” dissolution is a process that applies to many substances, not just salt, but it is still a process with a specific technical meaning (Shanahan & Shanahan, 2008, p. 53). Nominalization in history texts often stands for a process that takes place over a period of time and may be the historian's unique way of characterizing events, as illustrated in the sentence below:
- The enlargement of the nation's capacity to produce weapons, the advent of the aeroplane, and the improvement in worldwide communication systems through the telegraph increased the likelihood that the United States would enter the war. (Shanahan & Shanahan, 2008, p. 53).
The factual information in the sentence (e.g., that the nation produced weapons) is secondary to the interpretation of events, captured in the nominal forms—that more weapons (enlargement) and communication (improvement) increased the possibility (likelihood) of entering the war. Not only must the reader parse the basic subject/verb/object structure (enlargement, advent, and improvement increased the likelihood) but must also grasp that events that unfold over time, taken together, can influence other events. And, in Example 3, considerable post-modification follows each of these nominal forms. Taken together, the informational density and parsing constraints imposed by long nominal groups, whether the NP functions as grammatical subject or object, are substantial. Correct parsing requires defining the boundaries of the NP and determining head noun status. Elsewhere Scott (2004, 2009b) showed a clinical example of a student (10 years old) with a reading impairment who incorrectly concluded that mountains in the following sentence was the grammatical subject of the verb was divided, having incorrectly parsed the NP structure of the subject NP (shown in italics):
- The land to the west of the Appalachian mountains was divided into two territories.
Students who do not grasp the grammatical structure of nominal groups like these will be hampered in their ability to comprehend sentences, with the result that text level comprehension (i.e., gist) will be compromised as well. It is also unlikely that they will use NPs effectively in their own expository writing.
A second major way that sentences in academic texts are distinct from conversational discourse is the way that logical connections between clauses are marked. In conversational discourse with two or more participants, a small set of coordinating conjunctions predominate (and, but, so), in sentence or clause-initial position. The coordinator and is used particularly often with multifunctional meanings to signal a range of logical relations that include temporal sequence, consequence, comparison, and addition (Schleppegrell, 2001). Because text organization is coconstructed between conversational participants, and intonation can be marshaled to assist, conjunctions that form loose connections between clauses suffice. In contrast, academic texts draw on a larger set of conjunctions with more semantic specificity (Fang, 2006; Schleppegrell, 2004).
Clauses that relate logically are “loaded” into sentences in the form of restrictive and nonrestrictive relative clauses, object complement clauses, and adverbial clauses (illustrated below). Rather than a linear chaining (one clause added to another via coordinating conjunctions with equal syntactic status), multiclausal academic sentences integrate ideas by relating clauses one to the other in a tightly orchestrated hierarchical manner where one clause is subordinate to another (the independent-dependent relationship). Not infrequently, there are several levels of subordination within a single sentence (one clause is subordinate to another clause that is itself a subordinate clause). In the sentence below, a 12-year-old writer is summarizing a video that presented information on plant and animal ecosystems in desert climates:
- The plants adapt by making sure when there is water they get as much of it as possible by having wide arms or “tapping” for water. (Scott & Lane, 2008)
The hierarchical nature of the clauses in this sentence is illustrated below by specifying the dependency level of clausal verbs:
The plants adapt (main clause)
by making sure (first level of subordination via nonfinite adverbial clause telling how the plants adapt)
they get as much of it as possible (second level of subordination via object complement clause, i.e., the clausal object of the verb making sure)
when there is water; by having wide arms or “tapping” for water (third level of subordination via adverbial clauses that tell how plants will get as much (water) as possible)
Note that clauses with a lower level of dependency can actually precede clauses with a higher level of dependency, posing an additional processing challenge.
Science texts in particular utilize subordinate clauses in which the subject (or relative pronoun that stands in for the subject) has been omitted and the verb is a nonfinite form (i.e., not marked for tense or number). Examples are shown below in 6 and 7. The relative clause in Example 6 is missing that is at the point of the arrow, resulting in a reduced relative headed by the nonfinite verb returning:
- The blood ▴ returning from the body through the right side of the heart and to the lungs contains cellular waste. (Life Science, 2005, p. 542)
The reader might not realize that the verb contains is the main clause verb and belongs with blood rather than lungs. In Example 7, the adverbial clause is missing smaller dams are at the arrow.
- Smaller dams uproot fewer people and do less harm to the environment, while ▴ still providing energy for a region to grow. (Fang, 2006, p. 498).
Proper comprehension of this sentence requires that the reader calculate smaller dams as the subject of providing. These deletion patterns set up long-distance dependency relations that add processing complexity. In some cases, the subordinate conjunction is also omitted (if one reads Example 7 without the conjunction while, the sentence poses an even greater processing challenge).
Theme and information structuring patterns
A third feature of academic texts is the way that information is structured within and across sentences (here, we concentrate on mechanisms that operate within sentences). Linguists within the functionalist tradition (e.g., Halliday, 1994) identify this information structuring feature as the theme (the first structural element in a sentence that precedes the verb) and rheme distinction (the new information about the theme).2 The theme in academic texts is frequently the subject NP with a nominalized head noun, as was seen in Example 2, but the theme could also be other elements such as prepositional phrases or whole clauses. Related to theme/rheme structure is a mechanism known as “end-focus” (Quirk, Greenbaum, Leech, & Svartvik, 1985). Without intonation to highlight the new information in a sentence, written text frequently reserves the end of the sentence (the right-most part of the sentence) for new information. Adverbial subordinate clauses in academic tests often precede the main clause they modify to allow for the placement of new information at the end, as in:
- When an elephant drinks, it sucks as much as 2 gallons (7.5 liters) of water into its trunk at a time (National Geographic Kids, 2009)
As shown in Example 5 above, left-branching adverbial clauses (when there is water) can even be found at internal points within a multiclausal sentence. The placement of dependent clauses before the main clause subject and verb could place an additional load on sentence processing, as the listener/reader needs to “wait” for the most important information.
Combinations of the three strategies
For purposes of illustration, example sentences described thus far have isolated one of the three grammatical strategies. The reality, however, is that sentences in informational texts can become very long, containing all these structures as well as all three major types of subordination (i.e., adverbial, relative, and object complement clauses), as shown in the next example:
- Once DNA's structure was known scientists were able to figure out how it provides a library of instructions that control the cells that make up our bodies and those of other living things. (from a late elementary science trade book, cited in Fang, 2008, p. 477).
This 33-word sentence of five clauses begins with a left-branching adverbial clause (with a passive form of the verb was known) and continues with two object complement clauses and two relative clauses. There are four levels of subordination. The sentence concludes with a very long NP functioning as grammatical object of the verb provides, headed by the noun instructions and postmodified by two relative clauses. The focus (new information)—that scientists have figured out how DNA provides instructions to cells—must be extracted by the reader.
In sum, to explain the special grammar of informational text that older students must learn to comprehend and produce, we introduced three distinguishing characteristics that included field (what is talked about), tenor (the tendency to be authoritative), and mode (how material is organized). We then chose three grammatical strategies used as tools to accomplish field, tenor, and mode. Long and complex nominal groups help writers teach the material (field). Clausal subordination patterns, particularly adverbial clauses, convey authority by offering reasons, contrasts, and so forth (tenor). Subordination also contributes to text organization (mode), which is more hierarchical than linear. A third grammatical strategy of theme and information structuring also helps organize texts (mode) by assuring that the most important information is placed last.
CHALLENGES OF INFORMATIONAL GRAMMAR FOR OLDER STUDENTS WITH LANGUAGE IMPAIRMENTS
Facility with sentence grammar and meaning has played a central role in the study of older children and adolescents (C/A) with spoken and written language impairment (LI). Norm-referenced tests used to qualify children as LI for purposes of research participation or eligibility for intervention services typically contain sentence-level tasks. Sentences that are characteristic of academic text, with several clauses and long nominal structures, are found in tests designed for school-aged children (e.g., before you point to the house on the left side of the ball, point to the fish and the shoe). Because the names of tests/subtests draw attention to the task rather than the syntactic content (e.g., recalling sentences, following directions), when a C/A performs poorly, complex syntax may or may not be identified as a specific language weakness. Further, when a young elementary-aged child with LI becomes a middle or high-school student with a language-based learning disorder encompassing reading and writing, as is often the case (e.g., Catts, Bridges, Little, & Tomblin, 2008), difficulty comprehending and producing complex sentences has often taken a back seat to a focus on vocabulary and/or text-level features (Scott, 2009a). This is unfortunate, not only because older students must understand and produce such sentences “on their way” to comprehending and producing academic text, but also because they will have to deal with complex informational sentences at a metalinguistic level when revising their written work, learning correct punctuation, paraphrasing cited sources, and so forth.
Both direct and indirect evidence points to sentence-level difficulties in older children, adolescents, and young adults with LI when dealing with academic language. Direct evidence comes from studies that have employed sentence-level complexity measures. These are almost always investigations that have analyzed language samples. A typical finding is that participants do not produce complex (i.e., multiclausal) sentences at a comparable rate with age peers (Gillam & Johnston, 1992; Scott, 2003). There is also evidence that students with LI have more difficulty summarizing, recalling, and otherwise processing academic texts when comparisons are made with age peers (Ward-Lonergan, 2010a). These studies provide a type of indirect evidence to the extent that text processing requires sentence processing.
Several researchers provide direct evidence of difficulty with academic texts by using a design that affords a genre comparison (e.g., narrative vs. expository) or a modality comparison (speaking vs. writing) or both. Gillam and Johnston (1992) compared oral and written narrative language samples in children, 9 to 12 years of age. Their findings are relevant for academic texts because they provide insight into the impact of modality, and informational language is often written language. Children with language and reading impairments produced fewer complex sentences in writing compared to speaking—evidence of the special processing load imposed by the act of writing. By comparison, children with typical development used a higher proportion of complex sentences in writing. Of particular note, the types of sentences that best distinguished children with and without language/reading impairments were those that contained several different types of subordination (which, by definition, were sentences with at least three clauses). This finding underscores the fact that academic texts containing sentences like Example 9 above would be very challenging for students with LI.
In one of the few studies with genre and modality comparisons, several sentence measures distinguished children with language learning disabilities (LLD, mean age 11;6) from age peers (Scott & Windsor, 2000). Children gave both oral and written summaries of two educational videos, one that told a story about a young Greek boy (narrative), and another that described plant and animal adaptations and interdependencies in the desert (informational). Summing across genre and modality, grammatical complexity as measured by average sentence length was significantly lower for children with LLD, but did not differ significantly from language-matched peers (who were on average 2.5 years younger). Instances of grammatical error distinguished children with LLD from language peers, however. In a more fine-grained analysis of the same data, Scott (2003) reported that children with LLD had significantly lower rates of relative and other postmodifying clauses, object complement clauses, and coordinated clauses. A second analysis (Scott & Lane, 2008) revealed significantly less subordination depth (i.e., using more than one level of subordination). For all children, genre affected frequency of subordinate clause type, such that adverbials and relative clauses occurred at significantly higher rates in expository summaries. Scott and Lane also calculated sensitivity rates—the ability of sentence complexity measures to accurately identify LLD participants, and they did this for each of the four types of language. Important for our topic here, written expository text was the most sensitive context for identifying LLD. A recent finding (Nippold, Mansfield, Billow, & Tomblin, 2008) provides additional evidence of the sensitivity of informational text for identifying individuals with LI. Nippold et al. examined both conversational and expository (describing a favorite game) oral discourse in a large group of adolescents (average age of 13) who were participants in a longitudinal study of language. As kindergarteners, they had been identified as either language impaired (specific or nonspecific) or as typical language participants. Whereas sentence complexity measures failed to distinguish the groups in conversational samples, significant differences were found in the expository samples.
Ward-Lonergan (2010a) recently provided a comprehensive summary of the literature available on expository text production and comprehension by school-aged children and adolescents with LI and/or LD. Representative of the tasks and measures used in this body of research are the following:
- Answering questions about videotaped social studies lectures
- Verbally retelling social studies lectures presented on videotape
- Recalling main ideas after reading expository passages
- Detecting text inconsistency when reading expository passages, with and without cues
- Producing details of a text that are consistent with the text structure signaled in opening statements
- Use of text structure organization when writing expository text
As anticipated, participants in these studies who met criteria for LI or LD did not perform as well as age peers on these types of tasks. As noted above, these types of studies do not allow us to isolate the contribution sentence-level grammatical difficulties because lexical and text-level difficulties (e.g., difficulty processing inference) could also explain poor performance, but some contribution seems like a safe assumption.
APPLICATIONS OF INFORMATIONAL GRAMMAR IN ASSESSMENT AND INTERVENTION
When assessing the information-related grammatical needs of older school-aged students who struggle academically, speech-language pathologists (SLPs) can draw on multiple sources. In this section, we provide information from two of these: (a) analysis of performance in the curriculum itself, and (b) criterion-referenced measures including language samples and specially designed probes. Both methods offer opportunities to observe grammatical features.
Referencing the curriculum
By surveying the language demands of the specific disciplinary materials in a student's “backpack”—science texts, math worksheets, social studies websites, and so forth—the SLP can gain insight into comprehension challenges. Similarly, completed student assignments, papers, and tests can illuminate both comprehension and production difficulties in various subject areas. Such curriculum-based investigation is worthwhile to (a) establish a plan for developing the student's language skills to meet the demands of the curriculum and (b) measure progress relative to meaningful reference points.
The ultimate aim of intervention services is to support improved academic performance. To do so, it is important to have some appreciation for how ways of thinking and learning drive specific grammatical features. In the abstract, this may appear overwhelming. But from a practical standpoint, it is possible to make a good start by surveying the contents of a student's curriculum or work samples for the kinds of linguistic features summarized in Figure 1 (Additional resources on higher level grammar, including assessment and intervention, can be found in Supplemental Digital Content at http://links.lww.com/TLD/A1).
Selecting curricular subjects in which a student may be struggling, the SLP can review some subset of materials that are likely to present grammatical challenges because they contain or call for the use of features like those listed in Figure 1. This survey can help narrow the field of possibilities for more detailed assessment or for more focused goals.
It was argued in the first section that each academic content area exposes a student to both discipline-specific concepts and processes and to more subtle discipline-specific linguistic styles and conventions. This specificity of language to particular content implies that student performance may well fluctuate across academic domains. For this reason, it is helpful to collect samples from the content domains and school tasks that are of greatest concern and to observe student performance on both the linguistic structures that tend to differentiate expository language from other forms of discourse, and any specific constructions that are more frequently used within the relevant discipline. It is to be expected that student performance will vary across tasks depending upon the performance requirements (receptive vs. expressive) and modality (oral vs. written) (see, e.g., Scott, 1988). Consequently, it is helpful to examine performance on more than one task and in both oral and written modes, whenever possible, to capture specific performance strengths and needs.
In the next section, we review procedures that have been used to describe grammatical knowledge more fully and directly in expository/informational contexts, across oral and written modalities.
Informal, dynamic, and criterion-referenced measures
A detailed understanding of what a student knows and how this knowledge is put to use can be more informative for planning intervention than can normative comparisons. The assessment procedures described below are best described as “informal,” “dynamic,” or “criterion-referenced” because SLPs can expand on the procedures by developing their own stimuli for a specific purpose or individual, or even within a particular instructional moment.
Judgment tasks for assessing complex sentences
One way to assess syntactic comprehension is to present well-formed and ill-formed sentences and ask the student to indicate whether they are acceptable. Comprehension of some advanced language forms, such as passives and center-embedded relative clauses, can be evaluated using this type of judgment task. For example, Miller and Paul (1995) described presenting semantically acceptable and semantically odd sentences, and asking younger (6- to 9-year-old) children to point to a “silly” or “okay” picture. The task can be adapted for older students by asking them to respond verbally with “Okay” or “Not okay” when presented with more complex passives, such as those with expanded subject NP, or those within a larger discourse context (for instance, a science text where several abstract terms are being defined). A similar task could be designed for more advanced relative clauses, such as those that are center-embedded or have different subjects than the main clause.
A variation on the judgment task is to present two sentences containing the same elements (nouns, verbs, etc.) with differing syntactic structures, and to require a same/different meaning judgment. The Comprehensive Assessment of Spoken Language (CASL; Carrow-Woolfolk, 1999) Sentence Comprehension subtest utilizes this format, and it is amenable to investigating comprehension of most of the advanced syntactic structures we discuss here. The pairs of sentences on the CASL subtest often differ along more than one dimension, making it difficult to identify the source of comprehension errors. However, that very complication can be revealing of the more subtle issue of combinatorial fluency with the various structures and combinations contributing to sentence complexity. An advantage of using this type of task in a dynamic or criterion-referenced manner is that scoring and interpretation of errors need not be limited to assigning a score of correct or incorrect. In developing a specific task for any given individual, follow-up questions to probe comprehension errors should be added to clarify how the student interpreted the pairs of sentences.
Comprehension questions that probe syntactic knowledge
Another method for assessing comprehension involves presenting a stimulus sentence and asking a question focusing on some critical element. Miller and Paul (1995) provide this type of procedure for center-embedded relative clauses. A stimulus sentence such as, The boy who chased the cow was wearing a hat is followed by the question, Was the cow wearing a hat? (p. 161). A number of such items, divided among those which question the main clause subject (e.g., Was the boy wearing a hat?) and those questioning the relative clause subject (e.g., Was the cow wearing a hat?) can be developed to determine whether correct responses exceed chance levels.
Researchers continue to seek new and more specific methods for delineating the contribution of sentence-level grammar to comprehension in expository discourse. Yet access to “comprehension” remains elusive; insights into what a child understands when presented with complex linguistic stimuli are often haphazard or serendipitous. Some attempts are being made to more systematically sample understanding of complex sentences in discourse contexts through carefully designed tasks. Gillam, Fargo, and Robertson (2009) developed and tested a think-aloud task to test comprehension of expository text. They found that ability of fourth graders to paraphrase each sentence in a text passage was closely related to their accuracy in answering comprehension questions about the passage. They concluded that think-aloud data can be a useful supplement to norm-referenced assessment.
This has inspired us to create a similar paraphrasing task, using short passages of expository text edited to contain a variety of complex sentences. As each sentence is presented one at a time, students are asked to explain what they learned using their own words. If a response does not show retention of critical information in the subordinate clause, a follow-up question is used to probe whether the student comprehended that information. Student responses in this dynamic method of assessment can provide insight into sentence structure-related errors that might not be otherwise evident. For example, one 13-year-old with LI gave these responses as she listened to a passage about the history of passenger cars being read aloud, one sentence at a time.
Stimulus: They didn't hold many passengers, and because they were made out of wood, they were fire hazards.
Question: What did you learn about passenger cars here?
Response: They're fire hazards.
Follow-up: Why were they fire hazards?
Response: ‘Cause the cars were on the train.
In her initial response, the student essentially repeated the main clause, “they were fire hazards,” but did not include information from the subordinate clause—in this case an adverbial clause—“because they were made out of wood.” When asked specifically for the information from that subordinate clause, her response gave general information about passenger cars, but not the information that provided the causal link between the main and subordinate clauses. This particular student had used and shown comprehension of adverbial clauses starting with because on norm-referenced tests we had given, but her comprehension in this discourse-level task appeared limited. The sentence here was constructed with both coordination and subordination, and the subordinate clause was placed earlier in the sentence than the main clause (see earlier discussion of end-focus) and may thus have been more difficult to understand because of the increased memory and processing demands associated with the length and noncanonical sequencing of sentence elements. As shown with this example, an on-line series of comprehension checks can increase the likelihood that parsing errors would be detected.
Syntactic analysis of language samples
As with conversational and narrative discourse, a sample of an individual's expository discourse can provide a rich source of information about syntactic performance. Such samples can be readily available in the artifacts of the classroom, such as sample papers. However, it will sometimes be informative to elicit a sample to assess a particular feature or skill and to observe processing effort as well as the product.
Nippold, Hesketh, Duthie and Mansfield (2005) described a 4- to 5-minute structured oral interview designed to elicit an expository language sample. Students were asked to talk about a favorite game or sport and were then prompted to explain why it was their favorite and finally to describe the features, strategies, and procedures associated with it, given the following script (Nippold et al., 2005 p. 1052):
I am hoping to learn what people of different ages know about certain topics. There are no penalties for incorrect answers.
A. What is your favorite game or sport?
B. Why is [e.g., chess] your favorite game?
C. I'm not too familiar with the game of [chess], so I would like you to tell me all about it. For example, tell me what the goals are, and how many people may play a game. Also, tell me about the rules that players need to follow. Tell me everything you can think of about the game of [chess] so that someone who has never played before will know how to play.
D. Now I would like you to tell me what a player should do to win the game of [chess]. In other words, what are some key strategies that every good player should know?
This type of sample would be likely to stimulate language associated with the temporal and causal aspects of procedural discourse–-coordinating conjunctions such as and then, adverbs such as first, and subordinate adverbial conjunctions such as so that and when and if.
Scott and Windsor (2000) elicited written language samples from adolescents using a summarization task. The participants first watched a brief (approximately 12 min) informational video and were asked to write a summary of what they had learned. Notably, the stimulus video was carefully selected to have a recognizable and coherent theme and macrostructure, and therefore lent itself to organized summaries of the type that would motivate grammatical tools such as subordination and NP expansion.
A number of quantitative and qualitative analyses exist that can help establish levels of performance and pinpoint particular areas of need. Quantitatively, it is possible to summarize the overall length and complexity of sentences used in a sample with measures such as Mean Length T-Unit (MLTU) and a clause density metric. MLTU is calculated by dividing the total number of words in a sample by the total number of T-units (defined as a main clause and any subordinate clauses that are attached to it (Hunt, 1970). MLTU is sensitive to variables such as amount of subordination and NP expansion; however, MLTU does not indicate the relative proportions of features contributing to length (Scott, 2009b).
Clause density reflects the amount of subordination present in a sample. It is calculated by summing the main and subordinate clauses and dividing that by the number of T-units or main clauses (Hunt, 1970). Together, MLTU and clause density offer a means of quantifying the amount of sentence complexity present in a sample and both measures have been shown to distinguish children with language disorders from typically developing peers (Fey, Catts, Proctor-Williams, Tomblin, & Zhang, 2004; Nelson & Van Meter, 2007; Scott & Windsor, 2000), although differences have not been found consistently across all studies (see review in Scott, 2009b).
Once measures like MLTU and clause density are calculated, various types of complexity can be labeled to better understand how varied and sophisticated the syntax is. Some evidence exists that the language of children with language and reading impairments is marked by a decrease in combinatorial fluency, the ability to use multiple and varied clauses within a sentence. One quantitative measure that has the potential to reflect this more mature form of grammar is percentage complex correct T-units (Gillam & Johnston, 1992). This value is calculated by dividing the number of grammatically correct T-units containing one or more coordinating, subordinating, complementing, or relative clauses by the total number of correct T-units. Other measures have been proposed in an attempt to capture developmental changes in combinatorial fluency (e.g., Scott & Lane, 2008), and this remains an active area of investigation.
Qualitatively, it is possible to classify the types of complexity found in a sample and to observe combinations of complexity. Features to observe could include type of subordination (e.g., relative clauses, adverbial clauses, and object complement clauses), which could further include analyses of specific semantic-syntactic structures (meanings of adverbial conjunctions, variety and abstractness of verbs, etc.). The type of NP expansions and other syntactic tools for adding emphasis or tailoring a text to a particular content/audience (e.g., extensive postmodification or end-focus) can also be analyzed. The depth of subordination could also be noted, that is, whether in sentences with three or more clauses, one of the subordinate clauses is actually subordinate to another subordinate clause. If a student produces this type of sentence, it shows capability with the more hierarchical forms of clause combining described earlier as characteristic of informational language.
Numerous norm-referenced tests use sentence combining as a method for measuring syntax, but because these tests cover a wide range of developmental levels, there tend to be only a few items reflecting more advanced sentence features—not nearly enough to elicit the array of features that would be necessary to develop a program of intervention in that area. A typical sentence-combining task involves presenting two or more simple declarative sentences to be used in the creation of a single sentence that would presumably require some form of compounding of nouns or verbs, phrasal expansions, embedding, subordination, and so forth. The scoring procedures in norm-referenced tests usually take away points for simply coordinating clauses by use of the conjunctions and, but, and or, and award points for grammaticality, semantic completeness, and “grace” (creation of a coherent, sensible sentence). Careful task construction can make advanced syntactic structures more likely, but some structures cannot be obligated in sentence-combining tasks, because there may be more than one acceptable way to combine the stimulus sentences. Further assessment is called for in such cases.
To contrive more obligatory contexts for production of particular features, some researchers and test authors have developed variations on sentence combining such as story completion (Eisenberg, 2005) and parallel sentence production (Lombardino, Lieberman, & Brown, 2005). These methods might more appropriately be termed patterned elicitation, because a model of the target type of sentence is provided prior to presentation of a stimulus picture and/or set of words or sentences to combine. Another option would be to develop a set of five to ten sentence pairs to combine and ask the student to use particular adverbial conjunctions in the creation of a new sentence, indicating that the adverbial conjunction should be at the beginning of the sentence.
The response mode for these tasks could be either oral or written; in some cases, it would be wise to evaluate in both modes to compare whether skills are demonstrated in one modality or another. Scoring procedures could simply indicate correct versus incorrect responses, but could also reflect degrees of variation from correct (as is done on some of the norm-referenced tasks, such as the sentence-combining subtest of the Test of Written Language–3; Hammill & Larsen, 1996). For example, one might give one point for correct use of the adverbial conjunction, one for movement of the subordinate clause to the beginning of the sentence, and one for overall grammaticality.
Intervention targeting the grammar of information
Much of the guidance for teaching expository language skills comes from education literature on improving reading and writing (see reviews in Biancarosa & Snow, 2006; Graham & Perin, 2007). There are comparatively few such studies with individuals with language impairment, and the majority concern preschool children. Cirrin and Gillam (2008) recently searched 19 databases for outcome data on language interventions delivered by SLPs in school settings (K–5), where the majority of children and adolescents receive treatment. They found only two studies on syntax, and morphosyntax was the target in both of these studies. Intervention studies for sentence- and discourse-level syntax with school-aged children and adolescents with SLI are scarce. Our searches (Balthazar & Scott, 2007; Silliman & Scott, 2009) uncovered only a small group of studies that focus on teaching sentence complexity to individuals with LI.
Nonetheless, the existing literature suggests that a number of strategies and procedures hold promise for broad application across populations. As there is thought to be considerable overlap among groups labeled variously as language impaired and learning disabled (Catts & Kamhi, 2005), the methods recommended by the authors of large integrative reviews in the field of education provide a good starting point. In the discussion below, we examine several general principles for structuring intervention to support the growth of sentence-level skills needed to fully participate in informational language.
Principle 1. Grammatical instruction should involve both the modeling of forms within the context of expository discourse tasks and the explicit manipulation of forms in decontextualized, metalinguistic activities.
Although the emphasis in English education has shifted over the past three decades from structured, explicit grammar instruction to naturalistic, implicit modeling and back toward the middle, the accumulated evidence is clear: teaching grammar exclusively in the traditional lessons familiar to most grammar-school students is NOT an effective way to improve written expression (Graham & Perin, 2007) or reading comprehension (Moats, 2009). The syntactic skills necessary for successful participation in the informational language of school are gained both through exposure to informational texts and direct language instruction.
Principle 2. Introduce sentence structures and provide opportunities for practice in all modalities.
A second general principle is based on a fundamental assumption that intervention in one modality (e.g., writing) can influence performance in others (e.g., reading and listening), but that each modality presents different challenges and may require different solutions. Functionally, this means that progress should be measured as the ability to extend from one modality to another.
Principle 3. Provide organized, intensive, and repeated exposure to target patterns, accompanied by opportunities to practice.
Although there is little evidence to guide overall decisions about treatment intensity (i.e., number and length of sessions per week; Balthazar & Scott, 2007), available studies suggest that progress can be expected given 20 to 60 minutes of focused intervention per week (Ebbels, van der Lely, & Dockrell, 2007; Hirschman, 2000; Levy & Friedmann, 2009). There is also evidence that the frequency of use of complex sentences can be increased through “priming.” Gummersall and Strong (1999) demonstrated such a priming effect with typically developing children, who produced more subordinate clauses when retelling a story if, immediately before retelling, they heard the story a second time and repeated sentences with subordinate clauses. In another study, school-aged children with LI doubled their rate of production of relative clauses after taking part in conversations where clinicians produced the same structures at an increased rate (Johnson, Marinellie, Cetin, Marassa, & Correll, 1999).
Principle 4. Draw materials for treatment from a range of content areas/subjects within the student's own curriculum.
Finally, the raw materials for treatment should be drawn from the student's own curricular materials across content/subjects, so that discipline-specific examples are presented. For example, it would make sense to work on NP elaboration via nominalization and postmodifying relative clauses in science texts where these structures play a frequent and key role in conveying information.
Successful models and intervention activities
An active movement has evolved over the past two decades in education (specifically reading instruction) to strike a balance between “authenticity”–-what we refer to here as contextualized instruction–-and meta-awareness (Fielding & Pearson, 1994). In general education practices, there is a well-documented advantage to several methods of teaching grammatical structure that strengthen students’ metalinguistic awareness of target forms (Biancarosa & Snow, 2006; Graham & Perin, 2007). Evidence of the utility of decontextualized instruction combined with context-based strategies is emerging in the language intervention literature as well (Ebbels, et al., 2007; Hirschman, 2000; Katz & Carlisle, 2009; Levy & Friedmann, 2009; Scott & Balthazar, 2009).
Although there is agreement that some degree of direct instruction on language structure is beneficial, there is also a continuum of opinion regarding the amount of instruction necessary. At one end of the continuum, Hirschman (2000) improved the use of complex sentences of a group of third- and fourth-grade students with SLI by providing 55 half-hour sessions of metalinguistic training over a 12-month period. These students were reported to improve both oral and written use of sentence complexity to a significantly greater extent than a control group. Hirschman noted that the students with the poorest complex sentence usage appeared to benefit the most from the metalinguistic instruction. In the middle of the continuum, Levy and Friedmann (2009) spent a substantial period of time teaching complex visual analysis codes to a 12-year-old with SLI before using the codes to teach syntactic movement as found in relative clauses, wh-questions, and passive voice sentences. The intervention lasted 16 sessions over a period of six months. The treatment was deemed successful in that significant improvements were observed and maintained as late as 10 months posttreatment. At the far end of the continuum, Weaver, McNally, and Moermann (2001), have advocated teaching only as much terminology or analysis technique as is necessary in the moment, a model with which they have reported some success. Recently, we have found evidence that a 9-week program involving a mixture of contextualized and decontextualized treatment procedures produced improved performance on three types of complex sentences in three adolescents with SLI (Scott & Balthazar, 2009).
Regardless of the amount or length of training reported, it is clear that explicit grammatical instruction can be a valuable part of intervention. There is even a suggestion that it may be especially important to provide metalinguistic instruction for students with SLI, as a conscious bridge to the cognitive processes supporting complex language performance (Hirschman, 2000).
Metalinguistic training regarding complex sentence structure may be more easily accomplished by working in the written modality, where structures can be revisited multiple times to attack and discuss particular patterns or features. Explicit grammatical instruction utilizing sentence-combining, sentence completion, and sentence deconstruction methods have been shown to be effective (Fang, 2006; Graham & Perin, 2007). Developmentally, sentence-combining ability has been shown to be significantly associated with sentence complexity in a naturalistic story-writing task (Scott, Nelson, Andersen, & Zielinski, 2006). Studies with typically developing adolescents have repeatedly shown that sentence combining is an effective way to increase both quality and accuracy of writing (Graham & Perin; Saddler & Graham, 2005), and furthermore, that less skilled writers tend to benefit more from this strategy than do more skilled writers. A systematic review of the effects of sentence-combining instruction compared with more traditional grammar-teaching methods reported results that favored sentence combining (Andrews et al., 2006). Studies also have shown positive effects of sentence-combining practice on reading comprehension (Neville & Searles, 1985; Wilkinson & Patty, 1993).
Treatment sessions may move fluidly between contextualized and decontextualized tasks, as in the literature-based units described by Gillam and Ukrainetz (2007). In their narrative-based intervention, a book serves as the unifying theme for both comprehension and production activities, and instruction on macro- and microstructures involved in narrative. For instance, a particular passage might contain several relative clauses that a student has difficulty understanding. Observing that difficulty, the clinician could create a series of sentences of the same pattern and provide the student with focused practice deconstructing and combining.
This method holds promise for informational discourse as well. One can imagine using a particular written text as the unifying theme for teaching comprehension and production of expository macrostructures and for focused, decontextualized, metalinguistic instruction on particular sentence structures. Student assignments can be the basis for other types of extended intervention activities. For example, “minilessons” on sentence structure can be incorporated into the creation of a written product (Eisenberg, 2006; Nelson, 2010), by expanding sentences, combining sentences, and contrasting structures to highlight a particular feature and its meaning. Nelson (2010) described a recursive writing strategy and noted that students responded well to such activities at all phases of the writing process, from drafting to revising and editing. A study by Hirschman (2000) taught 44 children with SLI, 9 and 10 years of age, to decompose and reconstitute multiclause sentences in the context of reading fables. When comparing pre- and post spoken and written language samples, children in the experimental group made significant gains compared with those children assigned to a control group. Whether assignments involve reading or writing, there are abundant opportunities for systematically building metalinguistic and linguistic skills within meaningful contexts.
Ideally, grammar instruction should be integrated into instruction within the curriculum, and by classroom teachers in each content area (Fang, 2006, 2008). Grammatical interventions should not simply be isolated grammar lessons (Silliman & Scott, 2009). Learning researchers argue for increased attention to discipline-specific language for typically developing students (e.g., Weaver et al., 2001), and emphasize the importance of language in the content-area instruction of students with LD (Deshler, et al., 2001). This dual need–-for improved language instruction in the regular curriculum as well as for more specific grammatical instruction for students with LLDs–-results in the potential for involvement of the speech-language professional as a key member of the educational team. Such collaboration has been advocated as an essential means of improving literacy instruction, binding the teaching of literate language to the fundamentals of language competence (Moats, 2009; Sawyer, 2010).
Whereas teachers have the content knowledge and teaching methodology to deliver curricular instruction to groups of students, SLPs have the training to analyze language to diagnose problematic areas and to generate specialized instruction for particular learners. Combining these areas of expertise is likely to produce the best instruction (Sawyer, 2010; Schuele, 2009).
1. Andrews R., Torgerson C., Beverton S., Freeman A., Locke T., Graham L., et al. (2006). The effect of grammar teaching on writing development. British Educational Research Journal, 32, 39–55.
2. Bailey A. L., Butler F. A., Stevens R., Lord C. (2007). Further specifying the language demands of school. In Bailey A. L. (Ed.), The language demands of school: Putting academic English to the test (pp. 103–156). New Haven, CT: Yale University Press.
3. Balthazar C. H., Scott C. (2007). Intervention for syntax and morphology. In Kamhi A. G., Masterson J. J., Apel K. (Eds.) Clinical decision making in developmental language disorders (pp. 143–165). Baltimore: Paul H. Brookes.
4. Berman R., Nir B. (2010). The language of expository discourse across adolescence. In Nippold M., Scott C. (Eds.), Expository discourse in children, adolescents and adults: Development and disorders (pp. 99–122). New York: Psychology Press.
5. Biancarosa C., Snow C. E. (2006). Reading next—A vision for action and research in middle and high school literacy: A report to Carnegie Corporation of New York (2nd ed.) Washington, DC: Alliance for Excellent Education.
6. Biber D. (1986). Spoken and written textual dimensions in English: Resolving the contradictory findings. Language, 62, 384–414.
7. Biber D. (1988). Variations across speech and writing. Cambridge, MA: Cambridge University Press.
8. Carrow-Woolfolk E. (1999). Comprehensive assessment of spoken language. Circle Pines, MN: AGS Publishing.
9. Catts H., Kamhi A. (2005). Language and reading disabilities (2nd ed.) Boston: Pearson.
10. Catts H., Bridges M., Little T., Tomblin J. B. (2008). Reading achievement growth in children with language impairments. Journal of Speech, Language, and Hearing Research, 51, 1569–1579.
11. Chall J. (1983). Stages of reading development. New York, NY: McGraw-Hill.
12. Cirrin F., Gillam R. (2008). Language intervention practices for school-age children with spoken language disorders: A systematic review. Language, Speech, and Hearing Services in Schools. 39, 110–137.
13. Deshler D. D., Schumaker J. B., Lenz B. K., Bulgren J. A., Hock M. F., Knight J. et al. (2001). Ensuring content-area learning by secondary students with learning disabilities. Learning Disabilities Research & Practice, 16, 96–108.
14. Ebbels S., van der Lely H., Dockrell J. (2007). Intervention for verb argument structure in children with persistent SLI: A randomized control trial. Journal of speech, Language, and Hearing Research, 50, 1330–1349.
15. Eisenberg S. (2005). When conversation is not enough: Assessing infinitival complements through elicitation. American Journal of Speech-Language Pathology, 14, 92–106.
16. Eisenberg S. (2006). Grammar: How can I say that better? In: Ukrainetz T. A. (Ed.), Contextualized language intervention: Scaffolding PreK-12 literacy achievement (pp. 145–194).
17. Fang Z. (2004). Scientific literacy: A systemic functional linguistics perspective. Published online 1 December 2004 in Wiley InterScience (www.interscience.wiley.com
18. Fang Z. (2006). The language demands of science reading in middle school. International Journal of Science Education, 28(5), 491–520.
19. Fang Z. (2008). Helping students cope with the unique linguistic challenges of expository reading in intermediate grades. Journal of Adolescent and Adult Literacy, 51(6), 478–487.
20. Fey M., Catts H., Proctor-Williams K., Tomblin J. B., Zhang X. (2004). Oral and written story composition skills of children with language impairment. Journal of Speech, Language, and Hearing Research, 47, 1301–1318.
21. Fielding L. G., Pearson P. D. (1994). Synthesis of research: Reading comprehension: What Works. Educational Leadership, 51, 62–67.
22. Gillam S. L., Fargo J., Robertson K. (2009). Comprehension of expository text: Insights gained from think-aloud data. American Journal of Speech-language Pathology 18, 82–94.
23. Gillam R., Johnston J. (1992). Spoken and written language relationships in language/learning impaired and normally achieving school-age children. Journal of Speech and Hearing Research, 35, 1303–1315.
24. Gillam R., Ukrainetz T. (2007). Language intervention through literature-based units. In Ukrainetz T. (Ed.), Contextualized language intervention: Scaffolding preK-12 literacy achievement (pp. 59–94). Greenville, SC: Thinking Publications.
25. Graham S., Perin D. (2007). Writing next: Effective strategies to improve writing of adolescents in middle and high schools - A report to Carnegie Corporation of New York. Washington, DC: Alliance for Excellent Education.
26. Grigg W., Donahue P., Dion G. (2007). The nation's report card: 12th grade reading and mathematics 2005 (NCES 2007-468). Washington, DC: U.S. Department of Education, National Center for Educational Statistics.
27. Gummersall D., Strong C. (1999). Assessment of complex sentence production in a narrative context. Language, Speech, and Hearing Services in Schools, 30, 152–164.
28. Halliday M. A. K. (1987). Spoken and written modes of meaning. In Horowitz R., Samuels S. J. (Eds.) Comprehending oral and written language (pp. 55–82). San Diego, CA: Academic Press.
29. Halliday M. A. K. (1994). An introduction to functional grammar (2nd ed.). London: Edward Arnold.
30. Halliday M. A. K., Hasan R. (1976). Cohesion in English. London, Longman.
31. Hammill D. D., Larsen S. (1996) Test of written language—3. Austin, TX: Pro-Ed.
32. Hirschman M. (2000). Language repair via metalinguistic means. International Journal of Language and Communication Disorders, 35, 251–268.
33. Hunt K. W. (1970). Syntactic maturity in school children and adults. Monographs of the Society for Research in Child Development, 35 (1, Serial No. 134).
34. Johnson C., Marinellie S., Cetin P., Marassa L., Correll K. (1999). Facilitating a child's syntactic style during conversational language intervention. A paper presented at the annual meeting of the American Speech Language Hearing Association. San Francisco.
35. Katz L. A., Carlisle J. F. (2009). Teaching students with reading difficulties to be close readers: A feasibility study. Language, Speech and Hearing services in Schools, 40, 325–340.
36. Levy H., Friedmann N. (2009). Treatment of syntactic movement in syntactic SLI: A case study. First Language, 12(1), 15–45.
37. Life Science (2005). Circulation (pp. 540–549). National Geographic Society Education Division, New York, NY: Hill Companies, Inc.
38. Lombardino L. J., Lieberman R. J., Brown J. J. C. (2005). Assessment of literacy and language. San Antonio, TX: PsychCorp.
39. Miller J. F., Paul R. (1995). The clinical assessment of language comprehension. Baltimore: Paul H. Brookes.
40. Moats L. (2009). Knowledge foundations for teaching reading and spelling. Reading and Writing, 22, 370–399.
42. Nelson N. (2010). Language and literacy disorders: Infancy through adolescence. Boston: Allyn & Bacon.
43. Nelson N., Van Meter A. (2007). Measuring written language ability in original story probes. Reading and Writing Quarterly, 23(3), 287–309.
44. Neville D., Searls E. (1985). The effect of sentence-combining and kernel-identification training on the syntactic component of reading comprehension. Research in the Teaching of English, 19, 37–53.
45. Nippold M. A., Hesketh L. J., Duthie J. K., Mansfield T. (2005). Conversational versus expository discourse: A study of syntactic development in children, adolescents, and adults. Journal of Speech, Language, and Hearing Research, 48, 1048–1064.
46. Nippold M. A., Mansfield T. C., Billow J. L., Tomblin J. B. (2008). Expository discourse in adolescents with language impairments: Examining syntactic development. American Journal of Speech-Language Pathology, 17, 356–366.
47. Perle M., Moran R. (2005). NAEP 2004 trends in academic programs. Three decades of student performances (NCES 2005-464). Washington, DC: U.S. Department of Education, National Center for Educational Statistics.
48. Persky H. R., Daane M. C., Jin Y. (2003). The nation's report card: Writing 2002. (NCES 2003-529). Washington, DC: U.S. Department of Education, National Center for Education Statistics.
49. Quirk R., Greenbaum S., Leech G., Svartvik J. (1985). A comprehensive grammar of the English Language. London: Longman.
50. Saddler B., Graham S. (2005). The effects of peer-assisted sentence-combining instruction on the writing performance of more and less skilled young writers. Journal of Educational Psychology, 97, 43–54.
51. Sawyer D. (2010). Improving reading instruction: A call for interdisciplinary collaboration. Topics in Language Disorders, 30(1), 28–38.
52. Schleppegrell M. (2001). Linguistic features of the language of schooling . Linguistics and Education, 12(4), 431–459.
53. Schleppegrell M. (2004). The language of schooling: A functional linguistics perspective. Mahwah, NJ: Lawrence Erlbaum.
54. Scott C. (1988). Spoken and written syntax. In Nippold M. (Ed.), Later language development: Ages 9 through 19 (pp. 45–95). San Diego, CA: College Hill Press.
55. Scott C. (1994). A discourse continuum for school-age students: Impact of modality and genre. In Wallach G. P., Butler K.G. (Eds.), Language learning disabilities in school-age children and adolescents (pp. 219–252). New York, NY: Merrill.
56. Scott C. (2003). Literacy as variety: An analysis of clausal connectivity in spoken and written language of children with language learning disabilities. Paper presented at the 24th Annual Symposium on Research in Child Language disorders, Madison, WI.
57. Scott C. (2004). Syntactic ability in children and adolescents with language and learning disabilities. In Berman R., (Ed.), Language development across childhood and adolescence (pp. 111–134). Philadelphia: John Benjamins.
58. Scott C. (2009a). A case for the sentence in reading comprehension. Language, Speech, and Hearing Services in Schools, 40, 184–191.
59. Scott C. (2009b). Language-based assessment of written expression. In Troia G. (Ed.), Writing instruction and assessment for struggling writers: From theory to evidence-based principles (pp. 358–385). New York: Guilford Press.
60. Scott C., Balthazar C. (2009). Building complex sentences: An intervention feasibility study for school-age children with oral and written language disorders. Poster presented at the Symposium for Research in Child Language Disorders, Madison, WI.
61. Scott C., Lane S. (2008, June). Capturing sentence complexity in school-age children with/without language impairment: The search for more sensitive measures. A paper presented at the annual meeting of the Symposium for Research in Child Language Disorders. Madison, WI.
62. Scott C., Nelson N. W., Andersen S., Zielinski K. (2006). Development of written sentence combining skills in school-age children. A poster presented at the annual meeting of the American Speech Language Hearing Association, Miami, FL.
63. Scott C., Windsor J. (2000). General language performance measures in spoken and written narrative and expository discourse in school-age children with language learning disabilities. Journal of Speech, Language, and Hearing Research, 43, 324–339.
64. Shanahan T., Shanahan C. (2008). Teaching disciplinary literacy to adolescents: Rethinking content-area literacy. Harvard Educational Review, 78, 40–59.
65. Silliman E., Scott C. (2009). Research-based oral language intervention routes to the academic language of literacy (pp. 107–145–145). In Rosenfield S., Berninger V. (Eds.), Implementing evidence-based academic interventions in school settings (pp. 107–145). New York, NY: Oxford University Press.
66. Traxler M., Morris P., Seely P. (2002). Processing subject and object relative clauses: Evidence from eye movements. Journal of Memory and Language, 47(1), 69–90.
67. Ward-Lonergan J. (2010a). Expository discourse in school-age children and adolescents with language disorders: Nature of the problem. In Nippold M., Scott C. (Eds.), Expository discourse in children, adolescents and adults: Development and disorders (pp. 155–190). Psychology Press/Taylor & Francis.
68. Ward-Lonergan J. (2010b). Expository discourse intervention: Helping school-age children and adolescents with language disorders master the language of the curriculum. In Nippold M., Scott C. (Eds.), Expository discourse in children, adolescents and adults: Development and disorders (pp. 241–273). Psychology Press/Taylor & Francis.
69. Weaver C., McNally C., Moerman S. (2001). To grammar or not to grammar: That is not the question! Voices from the Middle, 8(3), 17–33.
70. Wilkinson P., Patty D. (1993). The effects of sentence combining on the reading comprehension of fourth grade students. Research in the Teaching of English, 27, 104–125.
1 We will use the terms informational tex}, academic text, and expository text interchangeably.
2 Other descriptors for theme/rheme distinction are topic/comment and given/new.