4.3 Understanding Text Difficulty

4.3 Understanding Text Difficulty

As the above discussion shows, texts are differentially complex because they "emerge from and realize the different purposes and contexts of language use in different situations" (Schleppegrell, 2001, p. 451). These complexities imply different ways of knowing and of learning. They present different degrees of challenge to the reader. Some lexicogrammatical patterns that contribute to text complexity are more familiar to students, whereas others are less familiar to them and hence pose greater comprehension difficulties. Thus, an important question to ask in reading instruction is this: what lexical and grammatical configurations of text complexity are potentially more challenging to students at a particular stage of literacy development? Instead of addressing this question, most discussions of text complexity for the purpose of reading instruction (e.g., Gamson, Lu & Eckert, 2013; Hiebert & Pearson, 2014) conflate text complexity with text difficulty, focusing on quantitative dimensions of text complexity as a key index in assessing the difficulty level of a text and in determining what texts students should be reading at different grades or language proficiency levels.

Commonly used quantitative measures — such as Fry, Flesh-Kincaid, Raygor, Lexile, or Coh-Metrix — employ either readability formulas or prediction equations. These measures have been widely criticized for their reliance on surface level linguistic features such as word length (the number of syllables per word), word frequency (the number of times a word appears), and sentence length (the number of words per sentence). According to Perera (1982), for example,

·　short words (e.g., ore) can sometimes be more difficult to read than longer words (e.g., hotdog),

·　words that are familiar to a reader with one meaning or in one grammatical context may become unfamiliar to the same reader when used with a different, less common meaning or function (e.g., the word "fault" in It's my fault. vs. How rocks move along a fault depends on how much friction there is between the opposite sides of the fault.),

·　sentences with the same vocabulary and of the same length can vary in difficulty because of the order in which the words appear (e.g., The menstrual cycle refers to the series of changes in the female reproductive system that includes producing an egg and preparing the uterus for receiving it. vs. The series of changes in the female reproductive system that includes producing an egg and preparing the uterus for receiving it is referred to as the menstrual cycle.), and

·　shorter sentences can sometimes be more difficult to comprehend than longer sentences due to the use of ellipsis (e.g., One kind of solidarity bee lives in holes in mud and another in holes in walls. vs. One kind of solidarity bee lives in holes in mud and another kind of solidarity bee lives in holes in walls.) or wording differences (e.g., We rely on forest vegetation for its life-sustaining capacity to generate oxygen through photosynthesis. vs. We need plants and trees because they can turn carbon dioxide into oxygen and if we didn't have oxygen we would die.).

As a result of these misjudgments, readability formulas can be highly unreliable. For example, according to Lexile, college-level books such as Fahrenheit 451 (Bradbury, 1953), The Stranger (Camus, 1942), and The Grapes of Wrath (Steinbeck, 1939) are at fourth grade reading level, and Captain Underpants Series (Pilkey, 1997-2019), an entertaining illustrated children's novel series by American author and illustrator Dav Pilkey, is more challenging than Of Mice and Men (Steinbeck, 1937), a novella more appropriate for young adults or adults.(https://www.daowen.com)

Additionally, Hiebert (2011) noted that readability formulas tend to inflate the difficulty level of informational text that repeatedly uses certain uncommon words central to its topic but underestimate the difficulty level of narrative fiction that uses an abundance of dialogue. Other scholars (e.g., Bunch, Walqui & Pearson, 2014; Graesser, McNamara & Louwerse, 2003; Templeton, Cain & Miller, 1981) have criticized readability formulas for failing to take into account other factors that also affect comprehension, such as conceptual difficulty, interest level, text structure, text cohesion and coherence, and reader's background knowledge and cognitive aptitude. For these reasons, Perera (1982) cautioned that a more thorough examination is needed of vocabulary, sentence structure, and other aspects of linguistic complexity not revealed by measures of word or sentence length (e.g., literary devices, idiomatic expressions), noting that "informed judgments by a thoughtful teacher may have advantages over the application of a readability formula" in selecting/writing reading materials and in planning reading instruction (p. 101).

A further criticism of readability formulas or prediction equations is that all of the quantitative measures give a single, summative measure of a text's difficulty in terms of grade level equivalent or a continuous score. Such "omnibus" measures do not tell teachers which part of the text is more or less challenging, as not all parts of a text are equally challenging; nor do they give teachers any insight into how the text should be used or taught. Even when two segments of the same text have the same readability score, the later segment may be less challenging because by the time the reader gets to it, she or he might have built some background knowledge for understanding, which may not be the case when reading the first segment.

Moreover, an emphasis on a single, static measure and the reader-text match can lead teachers to overlook texts that are above or below the readability score for a particular grade or proficiency level but may offer more relevant and timely content or have the potential to entice and engage students in reading and learning. In other words, an emphasis on readability scores and the so-called reader-text match in reading instruction has the potential to inhibit teachers from moving beyond pre-leveled texts to more relevant and timely resources that students may find endearing. It may also encourage the use of artificially engineered or simplified texts that are devoid of content and interest. In short, readability scores are of limited value to teachers in assessing text difficulty and in planning instruction with text. In light of this, Hiebert (2014) has suggested that the quantitative measures, though readily obtainable, should only be a starting point in the assessment of text difficulty, much in the same way that a number on a medical exam (e.g., blood pressure) is the starting point that triggers follow-up diagnosis or further evaluation.

To be fair, many proponents of quantitative measures are not unaware of the limitations of readability formula. They acknowledge the need to include as additional indicators of text complexity such qualitative dimensions as levels of meaning or purpose, structure, language conventionality and clarity, and knowledge demands. These qualitative measures are, however, not described with the same level of specificity and sophistication as the quantitative measures. In fact, the description is often provided in dichotomous terms, such as simple vs. complex, clear vs. ambiguous, explicit vs. implicit, conventional vs. unconventional, literal vs. figurative, single level vs. multiple levels, simple theme vs. complex theme, and low intertextuality vs. high intertextuality. Such a cursory account sounds rather general and vague, failing to pinpoint the exact sources of text complexity that likely engender challenges for reading comprehension. As noted earlier, not all complexities are equally challenging from the standpoint of comprehension (because of reader characteristics); and complexity does not necessarily correlate with difficulty. Sometimes greater complexity may actually make the text more comprehensible because simplified or engineered texts can obscure semantic relations and disrupt discursive flow in ways that make the text less coherent and thus more challenging to comprehend (Beck, Omanson & McKeown, 1982).

To summarize, current discussions of text complexity or text difficulty often fail to address such important questions as (a) which particular lexical, syntactic, and discursive features are likely to cause reading problems, (b) what is it about these features that make them potentially challenging to students, and (c) what is the rationale behind particular configurations of complexity. Texts are complex for functionally different reasons, and understanding these reasons is key to understanding text complexity and critical to developing advanced literacy and disciplinary literacies.