Bu çalışmanın amacı akademik söylemde edilgenleştirilen sözceleri anlamsal olarak 'edici' ve "etkilenen" rollerini üstlenen özne ve nesnelerin türleri açısından betimlemek ve genel olarak edilgenliştirmenin metinlerdeki sonuçlarını ortaya koymaktır. Bu amaçla, Dilbilim Araştırmaları Dergisi (DA) 2005 ve 2009 yıllarında yayınlanan makalelerden 10 tanesiyle bir veri tabanı oluşturulmuş ve edici ve etkilenen öğelerin türleri ve metinlerdeki dağılımları saptanmıştır. Çalışmanın sonuçları yazar araştırmacıların etken ve edilgen çatıdaki sözcelerde edici olarak soyut adlarla [+insan] özelliği gerektiren eylemleri seçtiklerini göstermektedir. Bu tür özneler çalışmada "sözde-edici" olarak adlandırılmıştır. Çalışmanın diğer bulgusu da edilgen sözcelerde yükseltilen nesnelerin çoğunlukla soyut adlardan oluştuğunu, daha az oranlarda ad tümceciklerinin "olgu" ve "eylem" aktarmak için kullanıldığını göstermektedir. Son olarak, dilbilim metinlerinde edilgenleştirmenin yaygın olarak iddia edildiği gibi nesnenin sözce başı konumuna getirilerek konulaştırıldığı durumların az olduğu, sözce başı konumunda daha çok yer, araç, amaç bildiren öbeklerin yer aldığı gözlenmiştir. Bu sonuçtan hareketle, edilgenleştirmenin konulaştırmadan çok bakış açısı yansıttığı ve edicilerden çok etkilenenleri önceleme işleviyle kullanıldığı ortaya çıkarılmıştır.

This study aims at describing the structure of passives in research articles in terms of the properties of agents and patients to see whether they present genre specific features. The corpus used in this study contains 9 research articles in Turkish from Dilbilim Araştırmaları, published in 2005 and 2009. The corpus used in this study consists of 10 research articles (RA) published in Dilbilim Araştırmaları Dergisi (DA) 2005 and 2009. Since all the articles are in linguistics, the results of the study are limited to this discipline. The findings show that a typical agent in active clauses is [-human] and [-animate]. There are two forms of agents in active clauses: Full nouns referring to the other researchers and the first person plural. Hence, all the agents are human subjects. The percentage of pseudo agents in active clauses is very significant (74%) when compared to the human agents (%21). This shows us a metaphorical feature of agency in RAs. The authors map the human features to abstract entities, which leads to a sort of personification. This can also be a sign of our perception of language as linguists: A living human like entity. Considering this great amount of their existence in RAs, pseudo agents appear to be representative feature related to passivization in scientific discourse of linguistic studies. Another finding is that patients, too, mostly refer to abstract entities unlike the patients of narratives.This can be either a property of a specific genre or specific fields, since most works in linguistics involve/deal with/investigate abstractions related to natural languages. The most remarkable result here is the high percentage of abstract entities. In Tomlin’s terminology, what is attentionally detected by the authors is the abstract terms related to the field of linguistics. We can claim that this finding can be a genre specific feature when compared to the typical agents and patients in narratives, or it can field specific. It is necessary to describe the patients in RAs of other disciplines such as economics, philosophy, education, etc. to show to conclude that this is a field specific indicator Passivization results in a change in the positions of the agent and the patient. Patients are said to be “promoted” since they are moved to sentence initial position and agents are demoted from subject to oblique. They may even be removed from the surface structure and left implicit when the agent is unspecified as in someone had stolen my wallet “cüzdanım çalınmış” or has [-human] being simply the causer as in the door was closed “kapı kapandı” (by the wind or someone). Since the most accessible entities in discourse are presented at subject position, grammatical subjects are considered to be the topic, and object to be the focus (Givon, 2002). Hence, toicalization is considered to be one of the functions of passive voice. Are patients in RAs promoted and topicalized? The distribution of the patients promoted and left at the pre-verbal position show that objects are not highly promoted. The number of the objects promoted to clause initial position is low when compared to those left in the pre-verbal or middle positions. Clause initial positions are mostly occupied by locatives in Turkish “Türkçede”, in this research “bu çalışmada”, in the example below “aşağıdaki örnekte”; instruments with the program X “X programı ile”; post positional phrases showing a purpose or a cause relation in order to obtain a critical analysis “metnin eleştirel çözümlemesi için”, for this reason “bunun için”; and long relative clauses modifying the subject. This indicates that the prominent function of passive clauses is not topicalization in RAs. Considering that the abstract objects express procedures or findings, the authors represent the parameters, variables, and the circumstances to ensure that the reader would not misinterpret the results or what is presented. The high amount of “authors as implicit agents” in general shows us that the authors try to hide themselves to create a discourse stance for scientific objectivity. However, this is not the only reason. Recall that reference to generic entities in claim structures also resembles a considerable amount. This increases the validity of the authors’ arguments. Therefore, the function of passives in such contexts seems to be hedging.