本研究旨在探討漢語副詞子句的語用功能。論文主要部份著重於口語會話
中表時間、條件、原因等副詞子句的分佈及用法。論文最後一部份將口語
語料的研究結果與書面語中的副詞子句的分佈做一比較。 本研究中
所使用的口語會話語料共有四小時， 以 Sacks,Schegloff, and
Jefferson(1974) 的言談分析 (conversation analysis) 模式為理論架
構。 並依照 C. C. Ford (1993)分類英語會話中副詞子句的方法將此中
文語料庫中的副詞子句依其主要修飾子句的相對位置及其語調形狀(
intonation contour)分為以下三類:(i)出現在主要子句前的副詞子句;
(ii)出現在主要子句後的副詞子句，其主要子句是一繼續的語調(
continuing intonation); (iii)出現在主要子句後的副詞子句，而其
主要子句是一結束的語調 (ending intonation)，以觀察在不同位置與語
調的副詞子句的言談功能。 研究結果顯示: 表時間與表條件子句傾向
於出現在主要子句之前，但表原因的子句則大多出現於主要子句後。此外
，絕大部份出現在主要子句後的表原因子句，其所修飾的子句是一結束語
調; 此表示說話者在說此類子句之前，已經把本來意識焦點中想說的話
說完了。換言之，大部份的表原因子句與其主要子句之間的關係極為鬆散
，幾乎是獨立存在的子句。這表示在語法上表原因的連接詞其實是個對等
連接詞(coordinate conjunc-tion)。 最後我們觀察書面語中副詞子句
的詞序，抽取天下雜誌、聯合文學雜誌中四十二篇文章，共十四萬字作為
分析的語料。一如口語語料，書面語中的表時間與條件的子句幾乎都出現
在主要子句前，然表原因的子句出現在前後的各佔一半。可見表原因的子
句之用法與其它類的副詞子句極為不同，它們在會話中常出現於帶有結束
語調之主要子句後，此乃口語的特色。 說話者偏好於使用表原因子句來
補充說明其前句話，或是回答對方的問題，以消除聽話者的疑惑，具有溝
通的功能。大致上，本研究顯示: 出現在主要子句前的副詞子句為引述下
文之功用，然出現在主要子句後的副詞子句是為補充解釋前面的句子。
通常出現在主要子句前的副詞子句其所修飾的範距較大。

The research reported here is intended as a contribution to an
understandingof the adverbial clauses, used by speakers in
spontaneous communication.This study, following Ford's model for
analyzing the adverbial clauses in herEnglish conversation
corpus (1993), aims at investigating adverbial clausesin
spoken Mandarin conversations on the basis of quantitative
analysis. Thereare four-hour conversationdatabase in this
dissertation research, consistingof 19 texts. The adverbial
clauses in the database were categorized into (i)preposed
clauses to their modified material across continuing intona-
tion,(ii) postposed clauses to their modified material across
continuing intonation, and (iii) postposed clauses to their
modified material across final intona-tion (rising question
intonation or final intonation). After an inspection ofthe
data, the findings suggest that thetemporal, conditional, and
concessiveclauses are favored to occur before their modified
material, but the causalclauses, after their associated
material. Causals appear as separated, intona-tionally
disconnected units far more often than do temporals,
conditionals,andconcessives. The data also show that causal
clauses are fundamentally differ-ent from temporal, conditional,
and concessive onesin their use. Especiallythe final causals
appear to serve a quitedifferent function, being more in
thenature of coordinate clauses rather than subordinate ones,
which comment on acause, relevant to the preceding clause. This
is related to the type of infor-mation they ususally introduce.
Causal conjunctions usually introduce back-grund, support,
and motivation for their associated material, while temporaland
conditional clauses are prototypically discourse linking and
framing intheir function--the former dealing with time and
the latter involving hypo-theticality. On the whole, my data
support Ford's claim that initial adverbialclauses form pivotal
points in the development of talk and present
explicitbackground for material that follows; on the other hand,
the adverbial clausesappearing after their associated clause
only complete a unit of informationwithout creating discourse-
level links or shifts.In contrast, when placedafter its
associated material but in continuous intonation, an
adverbialclause only presents new information elaborating the
associated clause. Yet,different from the former, whose
continuing intonation signals that there ismore to come and the
present utterance is still in progress, the adverbialclauses
following final intonation, though emerging through
conjunctions asextensions of previous units, also represent
separate and intonationally dis-connected units. Finally a
comparison of the placement of adverbial clauseswith respect to
their modified material between spoken and written data re-
veals that final causals in the spoken data outnumber ones in
the writtendata. We suggest that the preference for
causals occurring after theirmodified material in
conversation results from the interactional need.
The research reported here is intended as a contribution to an