Saturday, November 19, 2005

Me tarieth...

This is the third in a series of speculations about the "Me reweth" construction in this Middle English poem:

Nou goth sonne under wode --
Me reweth, Marie, thi faire rode.
Nou goth sonne under tre --
Me reweth, Marie, thi sone and the.

I previously mused on "Me reweth" as perhaps a middle voice use of the verb "rewen" (to rue), and upon learning of a category of verb use in Old and Middle English known as "the impersonal construction" (e.g., me reweth, him thynketh, us thyrsteth), I began to speculate on the possibility of an English middle-voice construction going back to Indo-European.

Before I get into that, I might ought (to use a Southernism, and what's the history of that?) to specify more clearly what the middle voice is. According to this site maintained by SIL (Summer Institute of Linguistics):

Middle voice is a voice that indicates that the subject is the actor and acts upon himself or herself reflexively, or for his or her own benefit. In the case of plural subjects, the actors may, perhaps, act upon each other.
That might not entirely clarify things. Let's try Wikipedia:

Some languages (e. g. Sanskrit and Classical Greek) have a middle voice. An intransitive verb that appears active but expresses a passive action characterizes the English middle voice. For example, in The casserole cooked in the oven, cooked is syntactically active but semantically passive, putting it in the middle voice. In Classical Greek, the middle voice is often reflexive, denoting that the subject acts on or for itself, such as "The boy washes himself." or "The boy washes."
Perhaps I can sum up and clarify. In Modern English, we have no middle voice construction, but the sense of a middle voice emerges in certain intransitive uses of verbs, e.g., "bathe," as in "I am bathing." The verb "bathe" can be used transitively, as in "Let's bathe the dog," where the verb "bathe" has the object "dog." With "I am bathing," however, no object occurs because none is needed. We could force this into the reflexive construction and say, "I am bathing myself," but we would only do this to clarify that we're not, for instance, bathing the dog. As this example and the two explanations above imply, the middle voice and the reflexive overlap and express similar meanings, namely, that the subject of the sentence performs an action that acts back upon the subject itself.

Now, let me return to my quest.

In thinking about the impersonal construction (e.g., "Me reweth") as perhaps a middle-voice construction in Old and Middle English, I decided to do a Google search, looking for my -- lo and behold:

Ruth Möhlig

(University of Cologne)

The Old English impersonal construction: a middle voice pattern?

Abstract (pdf)

Old English did not retain the Indo-European inflectional category of middle voice, which expresses a verbal action that is performed by the subject but refers back to it and affects it (Beekes 1995:239-42). Previous studies (e.g. Fraser 1985, Hermodsson 1952, Ogura 1990: 43-4) asserted that Old English expressed middle situations primarily by lexical means instead. Thus, intransitive verbs like steorfan 'to die', but also transitive verbs in ergative use, such as openian 'to become open', and reflexively used verbs like ahebban 'to raise oneself' have been identified as verbs with potential middle semantics. The present article, however, proposes arguments for a semantic interpretation by which the OE impersonal construction (e.g. me-ACC/DAT hyngria? 'Iam hungry', him-DAT scamode 'he was ashamed', me-ACC/DAT thynceth thaet ... 'it seems to me that ...') may be regarded as a grammaticalized pattern which specifically coded the middle voice. This interpretation is based on theoretical foundations laid by Cognitive and Construction Grammar (e.g. Goldberg 1995, 1996; Kemmer 1993, 1994; Langacker 1991, 1996, 2000), which acknowledge that not only lexemes but also grammatical patterns or constructions have semantic content, that grammatical patterns are prototypically structured categories, and that lexical meaning and constructional meaning interact. The article will also discuss how the subsequent loss of the impersonal construction in the course of Middle to Early Modern English affected the English voice system (Allen 1995, van der Gaaf 1904).
Now, that sounded interesting, even if difficult. So, I wrote to Ms. Möhlig, explained my quest, my query, my elusive quarry -- and she replied:

Thank you for your message and for your interest in my work -- which is still in progress. The paper you refer to was a preliminary version of my ideas presented to a small circle of German Scholars at the "Studientag Englisches Mittelalter 2003" in Potsdam, Germany. It is part of my doctoral thesis (Working title: "The impersonal construction in early English reconsidered: A study of grammatical meaning and grammatical change") which will hopefully be completed in October 2006. For this reason, I am very reluctant in handing out this paper to wider circles, as I am not allowed to publish parts of my doctoral thesis before I have actually submitted it. Plus, as I said, the ideas presented in this paper were preliminary and I have revised and altered quite a bit in the meantime. I hope you will understand this. Nevertheless, I will probably be able to give you some more information about the impersonal construction (IMPc) as you found it in the poem.

The IMPc is a marked, though not totally uncommon, syntactic pattern in Old English (it is certainly of Proto Germanic origin, as also e.g. German and Old Norse/Modern Icelandic have it). It is grammatically marked, because the verb is always in the 3rd singular form, regardless of the person and number of the first argument [i.e., the position occupied by a noun or noun phrase]. Furthermore, the pattern is marked because there is usually no nominative noun phrase which could function as a grammatical subject. There are cases where a second argument coded in the nominative is present, but then this noun typically is semantically non-agentive, i.e. it is no good candidate for subjecthood, either. It is also often in verb-second position (i.e. in "object position"). The IMPc occurs (with several formal variants) with about 60 verbs in Old English, some of which are found almost exclusively in the IMP pattern (e.g. hyngrian, thyrstan) and others with which this syntactic use is very rare (e.g. genealeacan, fremian). Ogura (1986) gives a very detailed account of the syntactic uses of the pattern and the verbs with which it occurs in OE and early ME. It has been proposed that the IMP verbs assign accusative or dative case to their first argument lexically. Thus, from a structural point of view, the IMPc should be lost around 1250, because this is about the time when lexical case marking (i.e. the ability of verbs to assign, e.g. dative or genitive case for a direct object or accusative or dative for a subject) seems to disappear in English (see e.g. Allen 1995 on this). Interestingly, however,the IMP pattern flourishes in exactly this period. It is found with about 120 verbs in Middle English, c. 70 of these are not inherited from Old English (i.e. are loan words or new coinings), or are inherited from Old English but are not found in IMP use there. This has set me to the question whether there are other factors but purely structural ones which influence the development of this grammatical pattern. These factors, I am convinced, are semantic and discourse-pragmatic, i.e. they have to do with the particular meaning(s) or function(s) of the IMPc in early English. One of its meaning components (in fact the prototypical sense of the IMPc) is grounded in the cognitive middle domain, which in turn is closely related to the reflexive domain (see Kemmer 1993 on this). In fact, many of the ME impersonal verbs show reflexive and IMP uses side by side (ME rewen 'to rue' is one of them). These verbs are often borrowings from French reflexive verbs (see also Ogura 1990, 1991), and it is commonly noted that the reflexive construction in French is -- at least in parts -- a pattern expressing middle situations. Thus, it seems as if the "middle sense" of the IMP pattern was still quite alive when these verbs were borrowed into English. However, it seems that towards the end of the ME period, IMP uses get more and more idiomatized and sometimes rather formulaic. In the EModE period (from c. 1450 on) you find only few verbs in fossilized IMP uses, e.g. meseemeth, methinks.

I don't know whether this short account can help you with your question. The verb ME rewen certainly belongs to that group of IMP verbs which show this "middle sense". I would suggest that the most straightforward literal translation of the lines in your poem into ModE is probably "I feel pity for your son and yourself" or "Your son and yourself arouse pity in me". This will, certainly, not do for a poetic translation but may hopefully give you some help in deciding on its deeper sense. I hope to have been of some assistance to you. I would be happy to learn on which translation you have decided in the end.
I'm speechless with gratitude. There must be a middle-verb construction to express my state. Perhaps: "Me overwhelmeth"?

Yes, Ms. Möhlig, your generosity has helped enormously. I now know far more about a verbal construction that until recently, I was totally unaware of -- or aware of only as an idiomatized expression such as the fossilized "methinks."

Me thynketh thaet me moste muche thenken, and thus lang here, me tarieth.


At 12:17 AM, Blogger Thomas said...

It is possible that the very high frequency of DAT-case subjects in older Germanic languages is explained via the gradual, but early, elimination of the middle voice paradigm.

However, you should keep in mind that most all dative-subjects express a state, whereas only a minority of Greek, Latin, or Sanskrit middles do the same. Therefore, the ultimate origin of DAT-subject constructions is likely to be different from the original function of the PIE middle voice.

At 3:50 AM, Blogger Horace Jeffery Hodges said...

Thomas, I've not thought about this in a long time, but I see your point.

I don't know much about this issue, but couldn't the differing relative frequencies be due to the fact that dative-subjects in English that express a state have survived (perhaps because they express an internal state of the speaker, a particular category?), and other dative-subjects have simply not survived?

By their survival, they would necessarily be the majority of dative-subjects in English even if they were only a minority of Greek, Latin, or Sanskrit middles.

Jeffery Hodges

* * *


Post a Comment

<< Home