2.1 Pawley&Syder’s Lexicalized Sentence Stem

Pawley and Syder（1983）propose the important concept of lexicalized sentence stems，or regular form-meaning pairings in the language.By a lexicalized sentence stem they mean“a unit of clause length or longer whose grammatical form and lexical content is wholly or largely fixed”（ibid：191192）.A sentence stem commonly consists of an expression which is something less than a complete sentence.Typically it contains three components：

（1）stem：nucleus of lexical and grammatical morphemes（usually consisting of the verb and certain arguments）

（2）inflection：realization of the variable constituents（labeled as category symbols such as NP and PRO）

（3）expansion：other optional constituents

For example，the following three sentences（［1］，［2］and［3］）are the conventional expressions of apology，from which a recurrent collocation can be abstracted with a grammatical frame：

［1］I’m sorry to keep you waiting.

［2］I’m sorry to have kept you waiting.

［3］Mr X is sorry to keep you waiting all this time.

Sentence stem：NP be-TENSE sorry to keep-TENSE you waiting

Likewise，a sentence stem can be abstracted from［4］and［5］—two sentences that are usually used to explain or justify conduct that is queried by reference to a distinguished precedent for our action：

［4］If it’s good enough for the Queen of England to wear jeans to work it’s good enough for me.

［5］If it was good enough for the Apostle Paul/Disraeli it’s good enough for me.

Sentence stem：If it be-TENSE good enough for NP，it be-TENSE good enough for me

According to Pawley&Syder，a sentence or sentence stem is lexicalized if it

（1）denotes a meaning which is culturally authorized，

（2）is recognized to be a standard expression for the meaning in question，and

（3）is an arbitrary choice.

For Pawley and Syder，two features of LSSs are particularly prominent.First，very few LSSs are absolutely lexicalized.Although we may list several defining features of LSSs，it is in fact not possible to make a clearcut distinction between what is lexicalized and what is not.Everything is just a matter of degree：

An expression may be more or less a standard designation for a concept，more or less clearly analysable into morphemes，more or less fixed in form，more or less capable of being transformed without change of meaning or status as a standard usage，and the concept denoted by the expression may be familiar and culturally recognized to varying degrees.（ibid：212）

A second important aspect of LSSs is that each LSS is socially and culturally recognized.They are not something only familiar to any individual speaker；they are familiar to the community as a whole：

［t］he expression is a conventional label for a conventional concept，a culturally standardized designation（term）for a socially recognized conceptual category.Rather than being a‘nonce form’，a spontaneous creation of the individual speaker，the usage bears the authority of regular and accepted use by members of the speech community.（ibid：209）

Pawley&Syder’s proposal of lexicalized sentence stem is a useful guideline to studies involving clause-level multiword expressions.LSSs and other memorized strings are seen to form the main building blocks of connected speech，and this repertoire of familiar sentence-length expressions is highly relevant to nativelike fluency in English.Pawley&Syder’s emphasis on social and cultural constraints within discourse communities seems to indicate that this concept will be of particular interest in professional settings.However，their argumentation suffers limitation in two aspects.First，although they have noted that the number of LSSs amounts to“at least several hundreds of thousands”（ibid：213），they have not specified any practical method to arrive at a relatively full list of all the sentence-length expressions familiar to the English speaker.Their approach is largely theory-driven，and there is no reason why some items are more typical than others.

Much of the empirical work can now be done with the help of corpora，and there are common assumptions shared by Pawley&Syder and Firthian linguists，such as Sinclair and his colleagues.For example，all of them regard language as a social phenomenon，recognize different degrees of lexicalization，and stress the crucial role played by phraseological units as a whole in serving communicative purposes.The corpus-driven approach provides a large amount of data，enabling a thorough investigation of such phrases in terms of frequency and formal features.

Second，Pawley&Syder have investigated LSSs mainly by examining spoken materials.In fact，there have been few corpus-driven studies of recurrent clause-level sequences used to express evaluative meanings in academic discourse.The present study is designed to fill this gap.