2.2. FSs and error analysis
2.2.2. Types of errors in the use of FSs
Two of the most well-known EA descriptive taxonomies are the surface strategy taxonomy and the linguistic category taxonomy. These taxonomies “focus only on observable, surface features of errors, a basis for subsequent explanation”
(Ellis, 1997b, p. 54). Each taxonomy has its own advantages and limitations which are discussed below.
2.2.2.1. The surface strategy taxonomy
As the name suggests, the surface strategy taxonomy “highlights the ways surface structures are altered” (Dulay, Burt, & Krashen, 1982, p. 150). It showcases how learners’ errors deviate from the target versions. In the surface strategy taxonomy, errors can be classified as omission, addition, misformation, and misordering (Dulay et al., 1982). James (1998) added another category called
26
blends to account for errors made by blending two grammatical forms together (See Table 2.5 for descriptions and examples).
Table 2.5. Summary of surface strategy taxonomy (Dulay et al., 1982; James, 1998)
Researchers often rely on these categories to develop their own error taxonomies that are appropriate for their studies. An example of applying the surface strategy taxonomy to analyze FS errors was provided by Mohammadi and Es-hagi (2018). They divided the FS errors they found in the argumentative essays written by 63 upper-intermediate Iranian EFL learners into 13 categories, including omission, addition, incorrect selection, elision, overgeneralization, overuse, approximation (using synonyms), circumlocution (paraphrasing), calque (literal translation), misuse, redundancy, priming, and conceptual transfer.
Although Mohammadi and Es-hagi’s (2018) error taxonomy did shed light on the strategies used by learners when they made errors in the use of FSs, the sheer number of categories complicated the analysis process. In addition, some categories in this taxonomy may be overlapping and confusing. For instance, the categories of misuse and incorrect selection appear to overlap. The category of approximation, by nature, covers various strategies (such as using superordinate terms). However, in their study, it was simplified into only using synonyms, leaving out other errors that would have been rightfully considered errors of
27
approximation. As a result, this error taxonomy was not adopted in the present study.
The surface strategy taxonomy has received sharp criticism. According to Dulay et al. (1982), the surface strategy taxonomy indicates the cognitive processes by which learners reconstruct the second language. Ellis (1997) considers this a doubtful claim as it assumes that learners “operate on the surface structures of the target language rather than create their own, unique structure” (p.
56). He argues that the value of the surface strategy taxonomy is unclear since it does not provide an accurate representation of mental processes. This might explain why the surface strategy taxonomy is not widely used in EA.
2.2.2.2. The linguistic category taxonomy
Another type of descriptive error taxonomy that has been used to analyze FS errors is the linguistic category taxonomy. It is based on linguistic categories (such as nouns, verbs, adjectives, and others), bearing a close association with the traditional EA practice for pedagogical purposes. The linguistic categories in this type of taxonomy align with those commonly used in syllabuses and textbooks. In the context of formulaic language research, Qi and Ding (2011) put forward an FS error taxonomy based on a linguistic category taxonomy (See Table 2.6).
Table 2.6. An error taxonomy for FSs (Qi & Ding, 2011, p. 170)
28
From their perspective, FS errors can be either internal or external. External errors occur when sequences are used in inappropriate contexts, while internal errors refer to errors in the components of the sequences. In the category of internal errors, errors are further divided into linguistic categories, including verbs, nouns, adjectives, adverbs, prepositions, articles, pronouns, and conjunctions. There is also a category for errors in more than one component of the sequences that would not fit well into the other categories. Qi and Ding’s (2011) FS error taxonomy embraces the essence of a linguistic category taxonomy. It provides a clear and straightforward method to classify FS errors. Moreover, as a type of linguistic category taxonomy, it allows for the quantification of errors (Ellis, 1997b), which is one of the objectives of the present study. For such reasons, this study adapted Qi and Ding’s (2011) taxonomy for classifying errors in the use of FSs.
Table 2.7 presents the final error taxonomy used in this study. Several adjustments were made to Qi and Ding’s (2011) taxonomy to make it more suitable.
Table 2.7. The error taxonomy used for the present study (Adapted from Qi and Ding, 2011)
First, the category preposition + article was removed. This category includes errors involving both a preposition and an article – two components in a
29
formulaic sequence. It appears to overlap with the category others which includes errors in more than one component as described above. Second, in light of the pilot study carried out before the main study, the categories of pronouns and conjunctions were also eliminated. The details pertaining to this decision and the pilot study are discussed in the Methodology chapter.