Chapter 16

Wikis > Chapter 16


In this blog, we will be discussing the differences that age can cause in language learning and how these differences can affect the mastery of a new language. According to Arnon and Christiansen, adults and children show differences in proficiency as well as time taken for mastery of language (Arnon and Christiansen, 2012). By examining the approaches and strategies of children learning their L1 and adults learning L2, we will look into the reasons why adults are rarely able to attain native proficiency when learning L2 and how age influenced such a result.

When discussing language acquisition, 2 of the mentioned theories are the critical period hypothesis (CPH), and the sensitive period hypothesis (SPH).  The Critical Period Hypothesis considers that there is a specific period of time (in terms of age) to learn and acquire a language. Lenneberg mentions that the CPH is a biologically constrained process with a specific timetable, in this case, the age. However, according to Bateson, rather than a specific age period, learning language is about the window of opportunities that open at different age periods. In this Wikichapter, we will be focusing more on the actual approaches, as well as building blocks used when learning language, rather than biological aspects and limitations.  


Approaches in Language Learning

Languages can be learnt in two manners, under naturalistic, or instructional settings. Previously, the settings determined the methods in which the students are taught. The methods were then called communicative method for naturalistic settings and grammar-translation method for instructional settings. Both methods are approaches to language teaching, whether is it under a natural environment or under supervision and structured curriculum.

Instructional Approach

The instructional approach was derived from the rote learning of grammar structure and applying the rules in translation of target language to native language and vice versa. The instructional approach was promoted due to its ability to increase intellectual development in the learners, as well as to learners to be able to write and read in the target language. The focus of the approach is about reading and writing the language, less of speaking and listening the language, as the motivation of learning a new language for the learners is the ability to read and write with understanding.

Previously named the “Grammar-Translation” method, learners were taught how to translate phrases and sentences, then large chunks of texts to ensure that the grammar structure of the native and target language is translated correctly. The focus on grammar and structure in the texts rather than its contents resulted in a limited mastery of the target language. In comparison, the learners were unable to create output spontaneously due to the lack of vocabulary mastery. With regards to such an outcome, Richards and Rodgers claim that “[T]hough it may be true to say that the Grammar-Translation Method is still widely practiced, it has no advocates. It is a method for which there is no theory. There is no literature that offers a rationale or justification for it or that attempts to relate it to issues in linguistics, psychology, or educational theory.” (Richards and Rodgers, 2008).

Considering both L1 and L2 learners, they will be able to identify specific rules and specific structures when texts are given to them. They will be able to identify the time aspect, the people involved from pronouns. However, they will fail to understand the actual content of the text due to the lack of vocabulary. The one-sided approach hence can cause various problems, due to the minimal vocabulary mastery and overbearing attention on grammar rules.

In conclusion, the instructional approach may be a useful method in drilling grammar and structure of the target language, but it does not provide learners with spontaneity and the creativity to speak and understand the target language. As such, the instructional approach will best be paired with other approaches to reach maximum mastery of the target language.

Natural Approach


The natural approach was first developed by Krashen and Terrell in 1970s and 1980s (Krashen and Terrell, 1983). There is more emphasis on communication rather than focusing on accurate grammar and sentence structure. There are three basic rules that Terrell stated, believing them to be the key of the approach.

The three basic rules are as follows:

1. The focus of instruction is always communication, rather than its form.

2. Speech production comes slowly and is never forced.

3. Early speech goes through the natural stages (yes/no responses, holophrastic, 2-word, phrases, sentences), similar to L1 learners.

As such, the natural approach is used as a contrast to the instructional approach.

Furthermore, the natural approach is used to reduce anxiety and more interesting situations in learning. From the shift in focus towards communication, students are less worried about accuracy and can have fun communicating to each other just by using the new language. There is also a shift in choosing vocabulary mastery over grammar mastery. Due to the new focus, it is essential to have a wider vocabulary database to be able to understand and converse.

From the approach, three visible stages can be observed in the learners: comprehension, early speech and speech emergence. Considering the focus on vocabulary, the crux of these stages is to make sure that the vocabulary becomes part of the learner’s long-term memory, a process that Terrell refers to as binding. Only through binding can the learners attain “speech emergence” stage.

While Terrell created the approach, Krashen provided hypotheses in the model to allow the approach to be used as an application to language teaching (Krashen and Terrell, 1983). The hypotheses are as follows:

The natural approach does not adhere to these hypotheses very strictly. While it is crucial to note that Krashen and Terrell came up with the approach, Terrell considers grammar to be beneficial in learning a new language (Krashen and Terrell, 1983) while Krashen’s monitor hypothesis emphasises that conscious learning has no effect on learner’s ability to generate novel language (Krashen and Terrell, 1983).

Children learners tend to use the naturalistic approach due to the fact that they learn on the go, surrounded by the language they are currently learning. On the other hand, adult learners may be learning naturalistically because of migration or they could be undertaking the instructional approach because of the lack of immersion and the language could be a foreign one in the learner’s country.

Looking at the above 2 methods, it is clear that most linguists favour the naturalistic, or communicative, approach for learning a language. However, studies have shown that the results of these 2 approaches are not so distinct from one another. Krashen, Long, and Scarcella claim that adult learners learning under the instructional approach actually master vocabulary quicker than children, and according to Clahsen and Felser, adults have trouble with morphological and syntactical processing(grammar). This is directly opposite from what the instructional approach concludes.

Building blocks of language learning

Despite the common knowledge that children learn language better than adults, there are in fact instances, or areas of language that adults are able to grasp quicker than children. In their paper investigating the cause, Arnon and Christiansen (2016) described 2 phenomenons namely undersegmentation and chunking. Undersegmentation occurs when a multiword sequence is mistakenly acquired as a unit, and only afterwards properly segmented into its individual words. Chunking occurs when words which are often used side by side gets fused together into a multiword unit. Although these may seem like “problems”, the following sections will explain how the 2 phenomenons come about, and how they might actually benefit learners.


Undersegmentation is caused by two main factors – linguistic environment (the input received) and lack of prior knowledge (awareness of wordhood).

Children-directed speech differs from adult-adult speech in many ways. On average, children-directed speech contains shorter sentences of 3.5 words compared to adult-adult speech, which have 7.5 words. Furthermore, up to 25% of children-directed speech are one word utterances. There is more repetition, and a more pronounced prosodic boundary. This difference in input causes the children to acquire words that are frequently uttered together, such as over here, show me, or want some more, as a single unit. Meanwhile, adults, who do not hear as much repetition or defined prosodic boundaries, and thus are less likely to perform undersegmentation.

Next, children learning their first language have no knowledge of what a word is, or that it even exists. All their learning come from the real time language that they are hearing, and is thus based on the prosodic boundaries, leading to undersegmenting when the words are uttered, and often, repeated, in multi-word prosodic units. On the other hand, adults are aware of word boundaries, and after undergoing literacy training in the midst of instructional learning, the exposure to the language’s writing will reduce the chances of undersegmenting words, allowing them to correctly segment words into individual lexical units. The evidence that the written form causes a lower rate of undersegmentation can be seen in how preliterate L2 learning children also has a high rate of undersegmentation like their peers learning L1. At this juncture, it is important to note that the children are in fact segmenting the speech they hear, but in the process extract units that contain more than 1 word.


Another process that will result in multiword units is chunking. Chunking had also previously been called serial recall and recognition (Miller, 1956). This process is more straight forward than undersegmentation. It derives from frequent exposure to multi-word phrases, such that they become “chunked” together in the mental lexicon. The process is complementary to language use as actual language is highly formulaic, with up to 50% of language used containing multiword expressions (Erman & Warren, 2000), that are in predictable, or restricted combinations.

Despite this, adult L2 learners often treat formulaic language much more flexibly than L1 learners do. This is in part due to the trouble that they have in identifying a chain of words that must be used as a fixed unit. Even after explicit instruction, L2 learners consistently deviate from the phrase they were tasked to remember (Wray & Fitzpatrick, 2008). On the other hand, children are learning both the words of their language and the way they are combined at the same time. Research has shown the eloquence with which children use statistical learning (Gervain & Werker, 2008). Thus, using words they already know, children will further segment neighbouring words, and by one year of age, have accumulated vast knowledge on the way their language is structures. This helps in their ability to recognize  “chunkable” phrases, and add to their repertory of formulaic language.In short, while children are able to make use of chunks rather naturally, adults appear to have difficulty recognizing and using chunks.


As mentioned in previous sections, naturalistic and instructional approaches will vary in their results for both the learning in children and adults. Basing languages purely on vocabulary and grammar, learners will pick up vocabulary better from naturalistic approaches while learners will learn grammar better through instructional approaches.

From the above sections, it is apparent that children use much more multiword units in the process of language learning than adults do. It is suggested that this will have an impact on their ability to learn grammatical and distributional relations between words, and subsequently, the way they learn multiword relations in L2. Due to this, the predictive relations between words are changed thus causing the differential learning outcomes between adults and children. This suggestion is based on 2 observations. Firstly, that prediction plays a key role in language, and secondly that first language acquisition does not happen in a developmental vacuum, meaning children are learning both form and function at the same time, unlike adults who already have both linguistic and conceptual representations in place due to their L1.

To illustrate the first observation, learning occurs “when there is a discrepancy between what is expected based on prior experience and what is encountered in the environment.” (Arnon & Christiansen, 2016, p.14). According to the discriminative learning models, prediction is not only  a driving force of learning, this learning will affect what is subsequently learnt. And informativity is the basis of prediction. Thus, seeing as adults who are less reliant on multiword building blocks have considerably more trouble with grammatical and distributional relations, they will have a lower starting point in this circular process of learning, and children would fare much better, as seen from Ellis & Sagarra’s (2010) conclusion that L2 cues that are learnt earlier are learnt better.

An example to illustrate the second observation is the way grammatical gender is learnt. When children learn their first language, they not making the connection that “pelota”(spanish for “ball”) means “ball”, like an L2 adult would be able to do. Children are at the same time learning everything about “ball” that L2 learners would already have a preconceived notion about. In this case, for L2 learners, “ball” refers to something round, probably bouncy, usually a toy. However, children L1 spanish learners who utilise multiword units would learn the article “la” together with “pelota”, forming “lapelota” in undersegmentation, or “la-pelota” in chunking. Thus, on top of these various meanings, “pelota” would also include the meaning of “female”, and this is engrained in their mind from the start. While having prior knowledge would be beneficial in the lexical mapping of individual words, adults will miss out on these grammatical relationships, as adding the grammatical gender on to “ball” will leave a lesser impression, as there would be no change in the core meaning that is already in the adult mind.


In conclusion, the teaching method definitely play an important part, with the communicative method overwhelmingly popular amongst linguists. That being said, reading and writing can only be learnt through education, thus, to gain fluency in speaking and written language, both methods must be used. Meanwhile, the instructional approach will reduce undersegmentation during language learning, teaching learners to correctly segment individual words. However, this might actually disadvantage learners, as multiword building blocks to language learning has been claimed to result in better acquisition of grammatical structure and distributional relations. At the moment, these predictions are in the initial stages of investigations, and because of the usage based nature of such research, quantitative comparisons must be made, and all future research in the area will be important to come to a comprehensive conclusion, whether in support or against this theory.


Arnon, I. & Christiansen, M. (2016). Building blocks of language learning. Under review.

Celce-Murcia, M. (2014). Teaching English as a second or foreign language (4th ed.). Boston: Heinle ELT.

Clahsen, H., & Felser, C. (2006). Grammatical processing in language learners. Applied Psycholinguistics, 27(01), 3-42.

Ellis, N. C., & Sagarra, N. (2010). The bounds of adult language acquisition. Studies in Second Language Acquisition, 32(04), 553-580.

Erman, B., & Warren, B. (2000). The idiom principle and the open choice principle. Text-Interdisciplinary Journal for the Study of Discourse, 20(1), 29-62.

Gervain, J., & Werker, J. F. (2008). How infant speech perception contributes to language acquisition. Language and Linguistics Compass, 2(6), 1149-1170.

Krashen, S. & Terrell, T. (1983). The natural approach. Oxford [Oxfordshire]: Pergamon Press.

Krashen, S. D., Long, M. A., & Scarcella, R. C. (1979). Age, rate and eventual attainment in second language acquisition. TESOL quarterly, 573-582.

Miller, G. A. (1994). The magical number seven, plus or minus two: Some limits on our capacity for processing information. Psychological review, 101(2), 343.

Richards, J. & Rodgers, T. (2008). Approaches and methods in language teaching. Cambridge: Cambridge University Press.

Wray, A., & Fitzpatrick, T. (2008). Why can’t you just leave it alone? Deviations from memorized language as. Phraseology in foreign language learning and teaching, 123.

by Neo Hui Ying, Zolene Koh Jie Ning, AY2016/17 Semester 1


Sub Wikis