4.2 Deconstructing Text Complexity
Text complexity is often equated with text difficulty as if complex texts are necessarily difficult to read and comprehend. For example, it is defined in the Common Core State Standards as "the inherent difficulty of reading and comprehending a text combined with consideration of reader and task variables" (NGA & CCSSO, 2010, p. 43). The complexity of a text is thus determined through a triad of features that include quantitative dimensions (e.g., word length, word frequency, sentence length, and text cohesion), qualitative dimensions (e.g., structure, knowledge demands, levels of meaning, and language conventionality and clarity), and reader considerations (e.g., motivation, knowledge, experience, purpose, and task sophistication).
There are several problems with this conception of text complexity. First, it conflates text complexity with text difficulty. Text complexity is an inherent property of text, and can be described in terms of its lexical, syntactic, and discursive features; whereas text difficulty refers to the degree of challenge that a text presents to its readers in terms of its conceptual, linguistic, and visual elements. Thus, although reader and task are important considerations in comprehension and hence in any discussion of text difficulty, they are technically not features of text complexity. In other words, while text difficulty can vary based on text characteristics, reader profiles, and task requirements, text complexity is a relatively more stable feature defined by the internal properties of text.
From a functional linguistics perspective, a text can be complex along two dimensions: lexical density and grammatical intricacy. Lexical density is the proportion of content-bearing words (e.g., nouns, verbs, adjectives, most adverbs) to the total discourse (Halliday, 1989). It can be calculated by dividing the total number of content-carrying words by the total number of clauses in the text. This measure determines the informational density of a text, with higher score meaning more information is packed into the clause. Grammatical intricacy, on the other hand, refers to patterns of syntactic organization. It can be measured by dividing the total number of clauses by the total number of sentences. The higher the index, the more grammatically intricate the text.
Presented below are three examples from Halliday (1989, p. 79), with content-bearing words bolded. Text 4-1A, which consists of 5 clauses, has a lexical density score of 1.6 and a grammatical intricacy score of 5. Text 4-1B, which consists of 3 clauses, has a lexical density score of 4 and a grammatical intricacy score of 3. Text 4-1C, which contains 1 clause, has a lexical density score of 12 and a grammatical intricacy score of 1.
Text 4-1A
You can control the trains this way, and if you do that, you can be quite sure that they'll run more safely and more quickly, no matter how bad the weather gets.(https://www.daowen.com)
Text 4-1B
If this method of control is used, trains will unquestionably be able to run more safely and faster, even when the weather conditions are most adverse.
Text 4-1C
The use of this method of controlling unquestionably leads to safer and faster train running in the most adverse weather conditions.
As can be seen from these examples, a text can become complex by being either lexically dense (i.e., packing a large number of content-carrying words into each clause through multiple modification and embedding, as well as use of adverbial and prepositional phrases) or grammatically intricate (i.e., building up elaborate sentences through paratactic and hypotactic relations). Different ways of meaning (e.g., speaking or writing) engender texts of different complexities (Halliday, 2005). In informal spoken contexts, complexity is accomplished through "prosodic structuring and clause chaining with conjunctions and discourse markers", whereas in written academic contexts it is accomplished through "hierarchical structure, lexicalization of logical links, and clause-linking strategies of condensation and embedding" (Schleppegrell, 2001, p. 454). Thus, it is erroneous to assume that spontaneous spoken texts are less complex than written academic texts. As Halliday (1989) has demonstrated, both types of texts are complex, but their complexities are manifested differently. As a general tendency, the more natural, spontaneous texts and texts that are oriented towards events and processes (e.g., personal genres such as recount and narrative) tend to have higher grammatical intricacy and lower lexical density. By contrast, the more planned, distilled texts (e.g., written texts of academic disciplines) and texts that are oriented towards things/ideas and their relations (e.g., factual and analytical genres such as explanation, report, and exposition) tend to have higher lexical density and lower grammatical intricacy.
Academic and disciplinary texts are often perceived to be grammatically more complex, structurally more elaborated, and semantically more explicit than everyday texts. Based on the results from large-scale corpus investigations, Biber and Gray (2010) challenged this stereotypical view, showing that academic texts and everyday texts are both complex but in different ways. Specifically, academic texts are more structurally compressed with phrasal (non-clausal) modifiers embedded in noun phrases, whereas elaborated structures with subordinate clauses (especially finite dependent clauses) are much more common in everyday texts (i.e., texts written for ordinary purposes). Because of its compressed discourse style characterized by a heavy reliance on nominal/phrasal structures, academic and disciplinary texts are much less explicit in meaning than alternative styles that employ elaborated clausal structures. This benefits disciplinary insiders who have the necessary background knowledge to fill the semantic gaps engendered by the compressed discourse style but disadvantages novice readers who may be used to the elaborated discourse style and are still in the process of expanding their linguistic repertoire.
Not only does text complexity vary along the "written-spoken" or "academic-everyday" continuum and by genre, it also differs across disciplines (Fang, 2012a; Fang & Schleppegrell, 2008). While texts in academic disciplines are generally lexically dense, they differ in the lexical and grammatical patternings that contribute to this density. For example, science texts often use (a) technical terminology to construct non-commonsensical interpretations of the world; (b) expanded noun phrases to pack a heavy load of information into the clause; (c) nominalizations to build technical taxonomies, distill information, and create discursive flow; and (d) logical metaphor to develop logical reasoning and advance argument. History texts, on the other hand, are often constructed in abstract language that infuses the historian's ideological perspectives. The abstraction is typically realized through the use of (a) generic nouns, which refer to groups of people, classes of things, documents, or institutions; (b) nominalizations, which turn a series of events, an action, a quality, or time sequences into "things" that allow historians to ascribe judgment and expand information; (c) evaluative words and phrases to indicate affect, judgment, and valuation; (d) a sandwich texture that juxtaposes narrative recount of historical events with abstract explanation and interpretation of these events; and (e) within-clause realization of causal relations that conflates causality and temporality. Mathematics texts typically draw on meaningful configurations of language, symbolism, and visual display to construe mathematical knowledge, processes, and reasoning. In English language arts, language patterns of all kinds are used in literary and non-literary texts, with the aim to entertain, divert, challenge, and educate readers. These patterns are motivated in that they enable disciplinary experts to present information, structure text, infuse perspective, and develop argument in ways that are consonant with the norms and values embraced by particular communities of practice. As Biber (1992) observed, text complexity is a multidimensional construct, with different types of structural elaboration reflecting different discourse functions and different kinds of texts exhibiting different complexity profiles.