In SLA research, one finding seems beyond reproach is that there is a set order in which students acquire grammar. This “internal syllabus” cannot be overridden and thus textbooks that present grammar unit by unit are pointless and worse ‘unnatural’, because students are unable to learn what is taught until they are developmentally ready.
The research that underpins these claims comes from three main sources. The first are the morpheme studies which attempt to emulate L1 research showing native speakers learn English morphemes in a fixed order. The second is Pienemann’s work which unlike the morpheme studies does not look at the order of acquisition of several forms but instead looks at the stages learners go through in acquiring a single form (questions for instance). The third are interlanguage studies.
Although this research is often discussed, I have found the details are rarely forthcoming. I was curious to know two things about these landmark studies. Firstly, how was ‘acquisition’ measured, and secondly what do they consider to be ‘grammar’? In this post I will be looking at the morpheme studies.
1. what falls under ‘grammar’ and what does not?
In the morpheme studies, a set of roughly 10 morphemes are usually researched. These vary slightly such as when researchers separate articles into ‘the’ and ‘a’, or look at long and short plural sounds but in general they don’t differ much between researchers. The list of morphemes include such things as plural forms (dogs), Copula (is) (He is happy), auxiliary be (he’s coming), irregular past, regular past, articles and possessive -s (John’s cat).
There is a lot that teachers would consider ‘grammar’ that is not included. For instance:
- I should play tennis. (Modals)
- If you like it, then buy it. (Conditionals)
- I’ve told you already (perfect forms)
- What are you doing? (Questions)
2. How is acquisition measured?
In the morpheme studies, a test subject is said to have acquired a grammatical form if they can produce it correctly in a test. The test, called a Bilingual Syntax Measure (BSM), is usually carried out on children and involves showing cartoon like pictures and eliciting language from the subjects. A researcher will, for instance, say ‘here is a girl, now there are two of them. So there are two _____?’ this is known as an ‘obligatory context’ as students have to use the correct form to answer.
The next stage is that researchers score the learners depending on whether they produce the correct form or not. For instance (Dulay and Burt 1974):
- totally correct ie. “she’s dancing” (2 points)
- half right, ie. “She’s dances” (1 point)
- wrong ie. “She’s dance” (0 points)
The scores of the entire group are then added up and plotted on a chart. The equation used was the sum of the whole group / the number of possible points x 100. If more learners correctly produce plural -s than produce possessive -s, then the researchers claim that plural -s is acquired before possessive -s.* In the morpheme studies a form was said to be ‘acquired’ if subjects produced it accurately when elicited 90% of the time.
What did they find?
Researchers seemed to find that all students acquired language in the ‘roughly’ the same order regardless of their L1. For instance Mitchell and Myles (2004: 43) argue that these results suggest ‘second language learners are guided by internal principles that are largely independent of their first language’.
The interesting thing to notice when looking at this table is that the orders found were not actually the same between researchers, which is a little surprising for a ‘universal’ order. That said some researchers seemed not to mind and grouped the morphemes into ‘sets’ which are acquired in order.
The eagle-eyed among you will perhaps spot that there are still some outliers here such as “articles” appearing in stage 2 yet 1st in Dulay and Burt and 11th in Hakuta.
Issues with this research
I was interested to discover, that despite the ‘Holy Grail of SLA research‘ status that the morpheme studies have achieved, they have been under scrutiny for almost as long as the have been around. Some of the criticisms levelled at this research is as follows (apologies for not being able to properly source the origin of these).
- morphemes with different meanings (a/the) were grouped together in some studies
- What was classed as ‘grammar’ was a very limited number of morphemes
- most of the early research was carried out on ESL learners, not EFL students
- The orders vary in different papers, notably Hakuta 1974 (n-1)
- students were all grouped together to obtain results, hiding individuals or national groups who may not have followed the “natural” order.
- the studies did not look at acquisition over time but rather just took a snapshot
- accuracy order does not necessarily mean acquisition order
- Students’ overuse of the target morphemes was not counted
- The “universal order” is more accurately thought of as the “Spanish student order”
(Note some of these criticism have merit and others less. Check Larsen-Freeman and Long 1991 for a more detailed explanation.)
Notably, the claim that the order is universal has started to look suspect as L1 does indeed seem to have some influence on L2 (something that will not surprise most teachers). Luk & Shirai (2009) have argued that researchers continue to promote the order as ‘universal’ ignoring the evidence that it seems to be affected by a students’ L1. Corpus research, for instance shows that students seem to acquire morphemes in a different order. For instance, Japanese has a possessive particle ‘の’ but no plural particle and Japanese students seem to learn possessive -s before they learn plural -S (Anecdotally, this chimes with my experience). Hakuta’s study had a similar results and interestingly, Hakuta found that articles, which do not exist in Japanese, were late acquired by the Japanese student he studied.
Luk & Shirai (2009) found that not only Japanese but Korean and Chinese learners (all of who lack plurals) generally acquired possessive -s earlier and both plurals and articles ‘later than is predicted’ by the ‘natural order’ hypothesis. Other authors have noted that salience (how easy it is to hear the morpheme in input) could also play a role in explaining the order. And another possible factor is frequency, which is ‘the second most popular of the suggested causes of the L2 functor acquisition order (after L1 transfer)’ (Goldschneider and DeKeyser 2002: 29)
So the ‘holy Grail’ seems to have a few cracks in it. One author who believes that the morpheme studies have been used to make claims that they could not support is Mike Swan, who notes:
We have no reason at all to believe that the learning of most grammatical items is constrained in this way: that for yet-to-be uncovered developmental reasons, students might need to learn comparatives before relative pronouns, dativizing verbs before quantifiers or infinitives of purpose before possessive ’s. To claim that learnability findings preclude the operation of a grammatical syllabus is a large and unjustified leap across a wide logical gap.(Swan 2018: 254)
*the research methods are actually a bit more complex than this and differed between researchers but I have simplified it for the purposes of the this post.