以病人為主問話強調蒐集病人的全盤性資訊:醫師的溝通技巧是促使病人主動發言的關鍵，而問候病人是醫病互動的開端。醫師使用適當的寒喧語和話題能促進病人發言。本研究目的在於檢視台灣醫師使用情境寒喧語與病方回應發言量之間的關係:情境寒喧語的哪些次類可引發病人較多的發言量引發病人較多發言量之情境寒喧語，其話題有何特點我們根據44個錄影語料分析醫師在問學開場階段中，對初診老年病人和病人陪同者(合稱病方)所使用的寒喧語。根據病方發言音節多寡9我們將寒喧語分為「低回應率寒喧語」(病方發言量小於1個音節)、「中回應率寒喧語」(病方發言量2-6.5個音節)和「高回應率寒喧語」(病方發言量20.5-67.5個音節)。其中，以高回應率寒喧語可引發的語意與資訊內容最豐富，醫病間有較多的問答互動(定位;grounding) ，所以我們又將「高回應率寒喧語」稱為「高延展寒喧語」。話題特性不同也是發言量不同的原因:高回應率寒喧語多屬於「社會關係話題」(例如詢問病人職業)，中回應率寒喧語屬於「問診準備談話」(例如詢問老年病人聽力狀況)。病人最熟悉自己的社會關係狀態，因此，此額話題可以引發較多發言量;而問診準備談話僅是醫療過程一環，所引發發言量有限。最後，我們認為醫師使用社會關係話題與老年病人寒喧不僅可以建立關係，病人所透露的訊息也是醫師在評估病人社會心理狀態的重要參考。

As in most social interactions， affective and referential goals are essential in doctor-patient communication: doctors need to establish a trusting relationship with patients so that they can collect holistic health information from them. Although good rapport should be established early in the greeting stage of the medical interview， exactly how doctors' greetings facilitate patient participation has not yet been fully explored. By examining 44 triadic medical encounters among doctors， elderly patients， and patient companions collected in a teaching hospital in southern Taiwan， we examine three research questions. Fir。f a， what greeting topics are commonly observed? Among these topics， which ones encourage more verbal participation from patients? And， how do they achieve this goal in terms of discourse structure and under the context of Taiwanese geriatric interaction? Three categories of doctors' greeting topics were identified according to the average number of syllables they elicited in response from patients' response: greetings with low response (eliciting 0.46 syllables)， mid-response greetings (eliciting 4.8 syllables)， and high-response greetings (eliciting 41.1 syllables) We find that lowand mid-response greetings dealing mainly with topics related to the preparatory sequences of the interview， such as seating arrangements and name exchanges (low-response)，and the patients' hearing ability or language background (mid-response). High-response greetings， such as those touching on a patient's occupation or medical history， are social-relational (Coupland 1994). These topics extend the conversation structure by developing subtopics and establishing‘grounding' (Paek and Horvitz 2000)， i.e.， background knowledge between doctors and patients. The patient is the expert on， and has exclusive access to， his/her social-relational aspects. We argue that it is the patient who holds the key to high participa tion and extensibility of topics. Other high-response greeting topics， such as the patient's age and the patient's companions， reflect the unique and complicated relationships between age and health status (Coupland et al. 1991， Giles et al. 1994) and between patients and caregivers. Since social-relational informati on is as important as biophysical information in evaluating a patient's health， we suggest that doctors facilitate the use of social-relational topics at the greeting stage. This will help doctors establish rapport with patients and encourage patient participation during the medical interview.