The goal for many language learners is to be able to communicate comprehensibly, effectively, and appropriately. For some, the goal may be to approximate as closely as possible the norms of native-speaker English. For others, this may not be a goal since they subscribe to the notion of English as an international language – English as it is used by people with no other common language and reflecting the cultural and linguistic identities of the people who use it. Despite learners’ personal goals for learning English, many will want their English to sound both fluent and natural, even if spoken with an accent that reflects their mother tongue. Many learners achieve a high level of fluency, yet are told that their English often sounds unnatural. What gives language the quality of sounding natural?

Multi-word chunks
There are many factors that can contribute to the naturalness of speech. One important factor is the extent to which the learners are using what are sometimes called multi-word “chunks,” as well as conversational routines or fixed expressions. Random patterns of words do not occur together in speech, but often occur as multi-word chunks. These may be two-, three-, four-, five-, or even six-word chunks.
O’Keeffe et al. (2007) give the following list of the most frequent sixword chunks that occur in the CANCODE corpus, a 5 million word corpus of spoken English:

Conversational routines and fixed expressions
A marked feature of conversational discourse is also the use of a subset of the multi-word units – conversational routines – that often have specific functions in conversation and give conversational discourse the quality of naturalness (Nattinger, 1980). These perform a variety of functions in spoken English and the teaching of these and other multi-word units is a feature of some recent English courses such as the Cambridge University Press series, Touchstone (McCarthey et al., 2006). Hence, when learners use English, in order for their usage to sound natural, utterances need to be expressed in the way they are conventionally said in English, and this is something that is often not possible to predict. For example, why do we say when we meet someone for the first time, “Nice to meet you,” and not “To meet you is nice”? Both have the same meaning but the former is said, not the latter. Our linguistic, or grammatical, competence provides the basis for creating many different ways of saying things, however only a small subset of possible utterances is ever actually said. Wardhaugh (cited in Richards, 1990) observes:
There are routines to help people establish themselves in certain positions: routines for taking off and hanging up coats; arrangements concerning where one is to sit or stand at a party or in a meeting; offers of hospitality, and so on. There are routines for beginnings and endings of conversations, for leading into topics, and for moving away from one topic to another. And there are routines for breaking up conversations, for leaving a party, and for dissolving a gathering . . . It is difficult to imagine how life could be lived without some routines.
Consider the following routines. Where might they occur? What might their
function be within those situations?

This one’s on me. I’ll be making a move then.
It was lovely to see you. I see what you mean.
Thanks for coming. Let me think about it.
I don’t believe a word of it. Just looking, thanks.
I don’t get the point. I’ll be with you in a minute.
You look great today. It doesn’t matter.
As I was saying . . .

In a classic paper on lexical routines, Pawley and Syder (1983) suggest that native speakers have a repertoire of thousands of routines, or “chunks,” like these. Their use in appropriate situations creates conversational discourse that sounds natural and native-like, so they have to be learned and used as fixed expressions. Research by Prodromou (cited in O’Keeffe et al., 2007) suggests that a key difference between the speech of advanced SUEs (successful users of English) and native speakers is the presence or absence of chunks. He also raises the issue of whether it is important or necessary for SUEs to set out to fully master the use of chunks, since they often mark membership of a cultural group (e.g., Americans) that learners may not wish to claim membership in. O’Keeffe et al., (2007) however, suggest that “students who do wish to push towards near-native fluency should be given appropriate exposure to and practice in the use of chunks” (p. 76).
Exposure to chunks can be achieved through:

Article from: Richards, J. C., ”Moving Beyond the Plateau From Intermediate to Advanced Levels in Language Learning”, 2008, Cambridge University Press.