THE LOSS OF THE CORRELATIVE SYSTEM OF EARLY ENGLISH AND ITS POTENTIAL IMPLICATIONS FOR THE OV - VO SYNTACTIC CHANGE

古英语使用指示词来追踪所指对象和连接从句;例如，（1）主句中的中性指示词pæs连接到下面的分句，后者以另一个中性指示词pæt开头：（1）& heo pa索纳pæs gefægnode，她立刻就说，GEN很高兴她有一整年的生计，然后她立即高兴起来，因为她有一整年的供应品。Pæs位于动词之前，因此提供了关于OV词序的关键线索，因为代名词宾语与完整的NP宾语不同，在古英语中很少发生外置。这个小句连接系统搭载在一个使用性别指示词的所指跟踪系统上，如（2）：（2）& pa sealde he heom mid his agenre handa ane trywene flascan，然后他亲手给了他们一个木制的酒瓶葡萄酒[.]。他们喝了酒，直到他们到了拉文纳里昂。 ”然后他用自己的手给了他们一个装满葡萄酒的木瓶子。他们喝了它（意思是：这个瓶子持续了他们），直到他们来到拉文纳。性别，不像格，从一个子句到下一个子句保持稳定，因为性别是名词的固有属性。（2）中的阴性指示代词pære指的是第一个分句的新信息，即阴性名词flascan“bottle”。给定的信息通常由人称代词（“话题连续性”）指代，而新的信息由指示词（“话题转移”）指代，如在（3）中，其中指示词pam成功地指示从旧话题hei到新引入的话题sum manj的话题转移。(3)hei sæde，pæt sum manj wære in pære ylcan mægoe，pamj/*i wæs他说有一个人在同一个家庭DEM是nama Martirius他说有一个人在同一个家庭，他的名字是Martirius。这种劳动分工在其他西日耳曼语中仍然存在，但在现代英语中却消失了。这种“独立”指示词在早期中古英语中逐渐减少（艾伦，2016; Los &货车Kemenade，2018），导致了现代英语中单数的“独立”this和that不再指代人的指称（Huddalle & Pullum，2002：1405-6），因此不再作为话题转换器使用。复数形式的these和those没有受到影响，这表明了与性别的丧失有关：与单数形式的指示词不同，复数形式的指示词在古英语中没有性别。假设的事件链如下：1）形态性别的丧失破坏了依赖指示词的小句连接系统。 2)这种小句连接系统，不再受到形态学的支持，突然从语言中消失（Lenker 2010），从而消除了语言OV特征的习得线索，加速了VO的兴起。 3)话题转移不再被支持，这对话语产生了深远的影响，模糊了“谁对谁做了什么”，因为现在只有人称代词可用于转移和继续的话题。 中古早期英语语序研究一直因缺乏关键时期的数据而受到阻碍。一个新的资源，早期中古英语的解析语言地图集（PLAEME; Truswell et al. 2019），现在已经可以填补这一数据缺口。

Old English made use of demonstratives for referent tracking and clause linking; the neuter demonstrative pæs in the main clause in (1), for instance, links to the following subclause, which starts with another neuter demonstrative, pæt: (1) & heo pa sona pæs gefægnode, pæt heo hæafde and she then at-once that-GEN rejoiced that she had ealles pæs gæres bigleofan all the year's livelihood 'and then she immediately rejoiced because she had an entire year's worth of supplies' (lit. 'and she immediatey rejoiced because of that, that she had an entire year's worth of supplies') Pæs precedes the verb and hence provides a crucial clue about OV-word order, as pronominal objects, unlike full NP-objects, rarely undergo extraposition in Old English. This system of clause linking piggy-backs on a referent tracking system that uses gendered demonstratives, as in (2): (2) & pa sealde he heom mid his agenre handa ane trywene flascan and then gave he them with his own hand a wooden bottle wines fulle [...]. Of pære hi druncon, oo pæt hi to rauennam of-wine full Of that they drank until-that they to Ravenna becomon. came 'and then he gave them with his own hand a wooden bottle full of wine [...]. They drank from it (meaning: this bottle lasted them) until they came to Ravenna.' Gender, unlike case, remains stable from one clause to the next, as gender is an inherent property of nouns. The feminine demonstrative pære in (2) refers back to the new information of the first clause, the feminine noun flascan 'bottle'. Given information is typically referred back to by personal pronouns ("Topic Continuity"), and new information by demonstratives ("Topic Shift"), as in (3), where the demonstrative pam successfully indicates Topic-Shift from the old topic hei to the newly-introduced topic sum manj. (3) hei sæde, pæt sum manj wære in pære ylcan mægoe, pamj/*i wæs he said that a man was in that same family DEM was nama Martirius 'He said that there was a man in that same family, whose name was Martirius'. This division of labour persists in the other West Germanic languages but is lost in Present-Day English. Such "stand-alone" demonstratives decline in Early Middle English (Allen 2016; Los & van Kemenade 2018), leading to the Present-Day English situation in which singular "stand-alone" this and that can no longer refer to human referents (Huddleston & Pullum 2002: 1405-6), and hence are no longer available as Topic Shifters. That the plurals these and those were not affected suggests a connection with the loss of gender: unlike the singular demonstratives, plural demonstratives were not gendered in Old English. The hypothesised chain of events is as follows: 1) The loss of morphological gender undermined a clause linking system that depended on demonstratives. 2) This clause linking system, no longer supported by the morphology, abruptly disappeared from the language (Lenker 2010) and hence removed clues in acquisition of the OV character of the language, accelerating the rise of VO. 3) Topic Shift was no longer supported, which had far-ranging consequences for the discourse, obscuring "who does what to whom" because now only personal pronouns are available for both shifted and continued topics. Word-order investigations in Early Middle English have always been hampered by a lack of data in the crucial period. A new resource, A Parsed Linguistic Atlas of Early Middle English (PLAEME; Truswell et al. 2019), has now become available to fill this data gap.