A "two-way relationship in English" revisited: on reciprocal expressions in early English, with a digression into Modern English uses.

ABSTRACT

The present study aims at providing a survey of the earliest English developments of reciprocal expressions, from the Old English discontinuous expressions, in which the elements are separated from each other by other sentence elements, to the present-day tight units like each other or one another. The data have been retrieved from several computerized text corpora. It comes out clearly that the reciprocal pronouns consist of two separate elements and their syntactic behaviour is regulated by such factors as the existence or non-existence of prepositions or other adverbial elements, word-order, function in the sentence and metre. The early expressions for reciprocity do not seem to share all the characteristics of the Modern English pronouns, nor are the same shade distinctions to be seen in Old and Middle English. The process of change in the uses and structuring of reciprocal expressions tan be seen as a grammaticalization process from units consisting of word sequencies to compound pronouns proper. The English reciprocals also illustrate a rare development in that it does not follow the general trends of language economy.

**********

Reciprocal pronouns in English are by no means an unexplored area of study. The purpose of the present short survey, however, is to present some theoretical approaches to what reciprocity means, give an overall survey of how it has been expressed in English throughout the ages and to briefly look at the development of these expressions as a grammaticalization process, that is to say as a development from expressions with several separate elements which also had different syntactic functions to the present-day system of compound pronouns.

Even though the study is based on a large computerized text corpus, it does hot take into account every single instance of each other and one another in the available diachronic corpora. The study thus illustrates a selective and somewhat limited corpus.

I am primarily interested in charting the earliest ways of expressing reciprocity in English, but the prescrit day usage may serve as a good starting point. According to Quirk et al. (1985), Present-Day English (PDE) has two reciprocal pronouns, one another and each other, and they are related to the reflexive pronouns in that they express a 'two-way reflexive relationship' (1985: 364-365). The following examples illustrate this relationship:

1) Adam and Eve blamed each other.

(cf. Adam and Eve blamed themselves).

They each blamed the other.

They are never weary of each other's company.

The twins wanted each other to be present at all rimes.

They talked one with another. / They talked with one another.

Dennis and Margot are in love (with each other).

Somewhat earlier, Curme in his Syntax (1931: 101), expressed the central characteristics of reflexives and reciprocals in the following way: "The reflexives refer to the subject of the proposition in which they stand, indicating that the action performed by the doer passes back to him, or is associated with him; thus the reflexive form can refer only to the subject of the proposition or clause in which it stands. The reciprocal pronouns, on the other hand, express mutual action or relation on the part of the persons indicated by the subject." This important difference can be seen for example in the two short sentences referring to Adam and Eve, quoted above in (1). Unlike the reflexive pronouns, the reciprocals can co-refer only to plural noun phrases (or noun phrases that have a plural quality), since reciprocity presupposes more than one participant. Biber et al. express this idea by saying: "As with reflexive pronouns, the reciprocal pronouns ... mark identity with the referent of a preceding noun phrase within the same clause, usually in subject position. They differ from reflexive pronouns in that the reference is to more than one entity and that there is a mutual relationship between the entities" (1999: 346). Each other and one another are both written as word sequences, but it is better to treat them as compound pronouns than as combinations of two pronouns. At the same time, they correspond to the correlative use of each ... other and one ... another in sentences like They each blamed the other. And: The passengers disembarked one after another. The pronouns, like the reflexives, cannot be used in subject position in finite clauses, whereas they could in the early phases of English, and they can co-refer only to plural noun phrases, since reciprocity, unlike reflexive expressions, presupposes more than one participant. But there does not seem to be such a constraint on reciprocals as subject in nonfinite verb clauses. We tan say 'The twins wanted each other to be present at all times.'

Quirk et al. (1985) argue that as to usage there is no difference in the use of the two pronouns each other and one another despite the prescriptivists' preference for each other for reference to two, and one another to more than two. Curme (1935) and others have indeed seen some tendencies in the use of the two phrases, but it seems indeed that no systematic distinction tan be discerned. There may, however, be some register-bound preferences, as Biber et al. have pointed out. According to their findings, one another is relatively common in fiction and academic prose, whereas each other is common in all registers (1999: 346).

Space does not allow a more detailed discussion on the concept of reciprocity, but some attention could be paid to an article by Dalrymple et al. (1998: 209), which deals with the issue in a more theoretical way. Two examples may illustrate one important aspect:

2) Many people at the party yesterday are married to each other.

Many people at the party yesterday have been married to each other.

Dalrymple et al. talk about the Strongest Meaning Hypothesis, which implies that the meaning of an expression is taken from a small inventory of meanings. We can predict which meaning the reciprocal will have in a given context: it will take on the strongest meaning that is consistent with known facts about the antecedent, the scope, and the context. In examples (2) we have to accept the fact that the sentence can be applied to a society where only monogamy is accepted!

One further study would deserve a brief comment. G6ran Kjellmer's article in ES in 1982 on the use of the English reciprocal pronouns each other and one another, extensively discusses, among other things, the acceptability of reciprocal clauses in different contexts. Like Quirk et al. (1985), he also notes, for example, that some syntactic functions are impossible for reciprocals. Passive constructions are such; they cannot function as genitival s-modifiers of the subject either. It was also noted earlier that reciprocals do not appear the in subject position, even though it is actually on record. Syntactically, according to Kjellmer (1982: 235), each other and one another tan function as:

                            Genitival                         Other
                     Head   s-modifier      Genitival     prepositional
                      of       of        of-modifier of    modifier of

Subject               -         -              -                -
Object                +         +              -                +
Adverbial phrase      +         +              -                +
Attributive phrase    +         +              -                +
Predic. Complement    +         +              +                +

On the whole, then, reciprocal pronouns occur only to a very limited extent as the head or modifier of the subject but more freely with other constituents, particularly with the predicative complement.

In his discussion on the semantic characteristics of the reciprocals, Kjellmer also emphasises the generally agreed plurality or 'non-singularity' of the antecedent noun. This obligatory plurality also emphasises the important difference between reflexive and reciprocal expressions in that the reciprocals express a two-way reflexive relationship (Porter 1953, Kjellmer 1982). Kjellmer makes one more allowance, which is that the plurality need not necessarily be overt; it is sufficient if the idea of plurality tan be recovered from the context (e.g. The husband may forestall tensions and uncertainties by confiding to his bride that lying in each other's arms will be bliss enough for these few hours; i.e. for them to lie).

Another characteristic of reciprocal expressions is symmetry, ideally perfect symmetry, but imperfect symmetry also occurs. Perfect symmetry implies mutual action or relationship. What is stated in a sentence is as true of A with regard to B as of B with regard to A, e.g., in a sentence like John and Mary kissed each other. Verbs like meet or resemble etc. are necessarily reciprocal in character and thus display perfect symmetry. Such verbs also often drop the reciprocal, even though not all such verbs share this tendency. Other verbs like Kiss, see, like, love can be used reciprocally, though they are not regularly so used. If there is not full symmetry in a reciprocal sentence, i.e., if what is said about A with regard to B is not equally true of B with regard to A, the sentence may be unacceptable. Some actions or relations are explicitly unidirectional, and symmetry is excluded a priori. But it seems that acceptability is a more-or-less feature so that the less symmetrical a reciprocal sentence is, the less acceptable it becomes. Both the linguistic expression and the reference of that expression have to be taken into account, and hence so also the pragmatic aspect. If the referent is a group or is seen as a mass without individual traits, somehow some notion of symmetry has to apply. But the demand for symmetry is doubtful, if we consider an example like The children followed each other into the church, quoted by Dalrymple et al. (1998: 194). There is no symmetry, but the context helps us to understand the meaning of the sentence.

From the point of view of Present-Day English the most interesting aspect of reciprocals seems to be whether there is a difference between each other and one another. Grammar books normally give them undifferentiated treatment, and it is often said that only prescriptivists used to see a difference in the reference to two or more. But Potter in the 50's, as well as Kruisinga and others see another difference in their meaning: "[E]ach other individualizes the members of the group, i.e. it considers them separately and suggests a mutual relationship of the same nature: love, affection, hatred, contempt, envy, interest, etc.--One another, as is suggested by the classifying indefinite article before other, considers the reciprocal relation from a general point of view.--One another, of course, does not exclusively occur in expressions and sentences of such a general meaning; it may, and generally does, refer to definite persons. But the reciprocal in which they are engaged is always looked upon as a particular instance or concrete manifestation of the general verbal idea in the abstract" (Kruisinga and Erades 1953: 535f.) But the distinction is very subtle and there does not seem to exist consensus about this very subtle distinction. Potter (1953) says, though, that he has round a growing tendency on the part of 'more sensitive speakers' to distinguish between each other and one another: "If the speaker is thinking of agents as individuals or single units, he will say each other; if he is thinking first of actions as shared or mutual, he will say one another."

Regardless of the above very delicate and difficult shade distinctions, it is possible to suggest though that there is a stylistic difference between the two pronouns. Both Kjellmer (1982: 252) and Quirk et al. (1985) refer to the usages found in the Brown corpus. First of all, reciprocals are a minority phenomenon all in all, only 159 instances in the whole Brown corpus, 114 occurrences of each other against 45 for one another Of course there are also other ways of expressing reciprocity, but when comparing the proportions of the two pronouns in each text category in the Brown corpus, if becomes clear that the highest proportions of each other are Press, Reviews, Skills and Hobbies, Miscellaneous, Fiction, Romance, Humour. The character of these categories is predominantly popular, and the language is fairly colloquial. There are only two categories in which one another represents half or more than half of the reciprocals, i.e. Religion and Learned and Scientific Writings respectively. From this we can perhaps even quite safely conclude that stylistically the two reciprocals are different and each other is the more common, everyday reciprocal and one another the rarer and stylistically more elevated one. Biber et al. (1999) had quite similar observations, as was mentioned before. They also comment on the fact that the reciprocals are so infrequent, saying that the distribution must be connected with the lower frequency in conversation; if there are fewer plural referents, there will also be fewer reciprocal pronouns.

From Kjellmer's extensive discussion of the reciprocals in Modern English, it can be concluded that three kinds of rules tan be distinguished for the use of each other and one another, i.e. syntactic (e.g. pronoun cannot take subject position), semantic and stylistic rules. In the semantic rules the most relevant characteristics are clear plurality of antecedents and perfect symmetry of actions. The stylistic rules demand that the choice of pronouns should be determined by the stylistic level of the text. Only a sentence that obeys all these three kinds of rules will be fully acceptable. But even these rules are observed in a hierarchical order. Syntax ranks highest, semantic rules second, and stylistic rules as lowest, but there are also intermediate shades, i.e. the phenomenon is not binary but there is a scale from completely acceptable to completely unacceptable through many intermediate shades. (1) But it seems that the Strongest Meaning Hypothesis is also successful in the analysis of the meaning and function of a reciprocal expression.

In PDE we clearly have two different compound units, and the relevant distinction in their use seems to be stylistic. Helena Raumolin-Brunberg (1997) has followed the development of the reciprocals in Early Modern English (EModE) and has seen the process of acquiring their compound character as a movement from discontinuity to unity (more on this below). For even earlier stages of the language, the extant material for the reciprocals is sparse, which makes it difficult to draw definite conclusions on their syntactic role. Some characteristics tan, however, be traced.

The roots of the discontinuity in the forms of the reciprocals, still obvious even in EModE, are however, to be round in Old English (OE). Bruce Mitchell (1985:116-118) discusses OE expressions for reciprocal relationship under eight headings:

(a) The personal pronoun 

--alone

--with a preposition 'between, among'

--reinforced by an adverb under Latin influence (e.g. gemaenelice)

(b) The personal pronoun + self: Alexandres aefterfylgendas ... hu hie hie selfe mid missenliean gefeohtum fordydon. (Or 142.10); and we magon us sylfe betwux us on life aele o[eth]rum fultumian to dam upplican life (AECHom ii. 356.12)

(c) O[thorn]er

(d) AEg[thorn]er, na[thorn]er, o[thorn]er ... o[thorn]er

--expresses reciprocity between two persons

(e) AEghwylc, aelc, gehwa, gehwylc, aenig, ... o[thorn]er

--reciprocity between more than two

(f) An ... o[thorn]er, *sure ... o[thorn]er

(g) Subject + adverb (e.g. togaedere)

(h) The verb alone (e.g. 'meet', 'kiss', 'love')

I am here mainly interested in the every/each paradigm. It is, however, to notice how closely related the reflexives and reciprocals are, as illustrated by the examples quoted in (b) above. Also we tan see that in OE there are different ways of expressing reciprocity between two persons and between more than two (points (d) and (e)). The following examples illustrate the OE types of expression: (2)

3) [thorn]aet an to o[thorn]ran ne si ge[thorn]eod. (Epit.Bened. 124.11; 'et ne unus ad alium coniungatur')

4) [thorn]aet an o[thorn]erne na derie. (Epit.Bened. 127, 29; 'ut unus alium non noceat')

5) amancg [thorn]am [thorn]e hi him an o[thorn]er betwynan spraecon. (Chrod.Reg. 99,31; 'cumque uicissim alique confabularentur')

6) And riht is [thorn]aet aelc cristen man eac o[thorn]erne lufie & healde raid rihte, & [thorn]aet aenig o[eth]rum ne beode butan ... Ne aenig ne syrwe ne o[thorn]rum ne swicie, ac healde aele o[thorn]erne mid rihte, & [thorn]aet oenig o[thorn]rum ne beode butan [thorn]aet he wylle [thorn]aet man him beode. Ne aenig ne syrwe ne obrum ne swicie, ac healde aelc o[thorn]erne mid rihtre getryw[eth]e. (HC O3 WulfHom 10c 207-8)

7) ac [thorn]aet hy symle gearowe syn, and geworhtum beacne hy butan etcunge arisende caflice gehwylc oberne forestaeppe and to [eth]am Godes weorce efste. (BenR 47.13)

8) ac [thorn] waere rihtost [thorn]te rinca gehwylc o[thorn]ru guide edlean on riht; (Met 27.25)

9) sceal [thorn]eah anra gehwylc o[eth]rum swican. (Rune 60)

10) and het gehwilcne oderne adwean fram fulum synnum. (AECHom ii 242.32)

11) for[eth]an de gehwa hoet o[eth]erne, na hyne sylfne. (AEgram 125.8)

12) For[thorn]on eal swa teofanade, se be teala cu[thorn]e, aeghwylc wi[thorn] o[thorn]rum; (OrW 44)

13) [thorn]aet eower sceal ceghwylc o[eth]res let [thorn]wean. (Hom Assmann 159.182)

14) and ofer eall hlydende aelc man cwae[eth] to o[thorn]rum. (AELS 23.617)

15) And hi micclum ege him ondredon, and cwaedon celc to o[eth]rum. (Mk(WS) 4.41)

16) for[eth]am ic wat [thorn]aet ure lufede aelc o[eth]erne swa micle swi[eth]or swa ... (Solil 44.5)

17) & heora [thorn]aer oeg[eth]er o[eth]erne ofslog. (HC O2 Oros 68; 'two')

18) also [thorn]aet hi lagon [eth]aer begen, heora aeg[thorn]er o[eth]res slaga. (AELS 25.587)

19) aeg[eth]er [thorn]ara eorla o[eth]rum trymede heoforices hyht, ... (HC OX/3 And 33; 'many')

20) Syb waes gemaene bam pam gebro[eth]rum, blis edniwe. AEghwaeder o[eth]erne earme betehte, cyston hie on clypton. (HC OX/3 And 31)

21) Ne aenig man oberne baeftan ne taele ne hyrwe to swyde. (HC O3 Wulf10c 205)

22) [THORN]a reaferas sippan ofsloh aelc o[eth]erne. (Hept 7.22)

23) Hi begunnon to sleanne aelc heora o[eth]erne. (Hept 74.1250)

The usage was clearly not established in OE, which is also confirmed by the variety of these expressions: 75 instances of aelc have been recorded for the present study, 10 of aeghwilc and gehwilc each, three of gehwa, and none of the infrequent word aeghwa 'every'. AElc, e.g., appears in three syntactic structures independent aelc alone (15) followed by other; also a few instances of aelc as a qualifier (14) to a noun, and with a genitive modifier (16, 23), usually a pronoun, of the type 'may each of us love the other'. For gehwilc (7-10) and aeghwilc (12-13) different syntactic structures have been recorded in the corpus, but in none of them dependent use of the indefinite. The rarity ofgehwa (11) and aeghwa in OE in general may explain their rarity in reciprocal uses.

For example Schrader (1887: 46), when commenting on AElfric's language, says that "[i]n aelc-o[eth]er ist der erste bestandteil immer subject, der zweite naheres oder entferntes object und demgemass immer flectiert. Die verbindungen verlangen das verb immer im singular". This claim holds true throughout the corpus with minor qualifications. The only exception to the rule, i.e. a singular verb, appears in AElfric (22) in my data. In some instances the reciprocal expression is in apposition to a plural subject and mostly then the verb also appears in plural, as in (23). (3)

In addition to the OE words with the primary meaning 'every'/'each', aeg[eth]er and aeghwae[eth]er also appear in reciprocal expressions, as noted by Mitchell. Examples of these are illustrated in (17-20). Some indication of the continuum from OE to the PDE usage tan be traced in (5), where a preposition has been added to the reciprocal phrase. A variety of forms appear in (6), thus illustrating the flexible use of reciprocals in OE.

We can sum up the OE situation by saying that one another still waits for its appearance until late Middle English (ME) or EModE. The every/each element of the reciprocal can function, as it in many cases does, contrary to PDE usage, as the subject of the expression, and we cannot yet speak about compound pronouns in the modern sense, even though in very few instances one might like to see some signs of such a structure, as in (22). An even less probable candidate is (9) where the elements clearly appear in different syntactic functions and the sequence is only due to metrical demands.

When moving to the ME period, along the continuum from OE, some new gradual developments can be traced. The ME uses have not attracted very extensive study, apart from a detailed description in an unpublished dissertation by Ding-Taou Sheen (see also the discussion on EModE uses below). His corpus consists of as many as 340 instances from a very large ME corpus. This, however, also includes examples of 'each' and 'other' that are not used reciprocally. He carefully lists all possible structures and combinations recorded in his data (Sheen 1988). (4) Examples recorded from the present corpus:

24) 'Hit is fayrer dethe to sley eche o[thorn]yr pen to 3eld vs to more payne.' And so xj M slov euyreche odyr and fel ded for stynche of dede men. (HC, M4 Siege of Jerusalem 86)

25) and perfore we schulde loue euery o[thorn]ere. (HC, M3/4, Vices4, 113)

26) bot if that on that other waste. (Gower CA Prol. 649)

27) hov holde 3e so silence: [thorn]at neuer on ne spekez with o[thorn]ur. (Seng. Leg. 228, 324)

28) Neuer one ano[thorn]er habade. ("Sir Perceval of Galles" in ME Metrical Romances, 549)

29) When men in er[thorn] ben most besi, [THORN]at neuer on o[thorn]er help [thorn]ai schal. (Poems of John Audelay)

30) Ne og ur non o[thorn]er to sunen, oc eurilc luuen o[thorn]er, also he were his broder, ... (HC M2 Best 12)

31) Many dou3tty 3onge kni3th; bat ilk day assayed his mi3th; Vche on o[thorn]ere, wip grete mayn, ... (HC M2 Kalex I, 53)

32) fif hundret italt of wepmen & of wimmen an hundret & [thorn]ritti [thorn]rungen euchan bitioren o[thorn]er forte beo bihefdet. (HC M1 Jul 121)

33) & se feste ilimet wi[thorn] lim of anred luue euch ofow to o[thorn]er. (HC MI Ancr 117)

34) Euch a[thorn]rusme[thorn] o[thorn]er. & euch is o[thorn]res pine. Ant euchan heated o[thorn]er. & himseoluen as [thorn]e blake deouel. (HC M1 Sawles Warde 172)

35) Armed men of hem ther sprong, And echon on other faste dong. (The Laud Troy Book, 32)

36) Faderr, sune, Hali3 Gast, Illc an effninng wipp o[thorn]err, ... (Orm 10991)

Mustanoja (1960:153-154) mentions that there were several ways of expressing reciprocal relations in ME, including active verbs like kiss alone, but usually reciprocity was expressed by 1. each ... other, every ... other; 2. one ... other, or 3. Either ... other, and their variants. The second type is according to him a ME innovation (sec example (26)). But Rissanen (1967: 97), in his dissertation on one mentions three very early instances (see examples (3)-(5) above) of one...other clearly according to him, formed as literal translations of the Latin original text. That seems acceptable. Either appears in reciprocal expressions as early as in OE (sec examples (17)-(20) above). No instances have been recorded of one ... another in the ME part of the Helsinki Corpus (HC) which seems to take this structure to the Early Modern period. Sheen quotes (1988: 161), however, example (28) where we have one another, but he also notes that even though the reciprocal expression is placed in the subject position, one is understood as the subject, and another, as object of the main verb habade 'waited'. Type every ... other (a compound unit or discontinued type) has been attested in the early ME period: We have an example of a discontinued pronominal phrase in (30) and somewhat later ones in (24) and (25). I would feel tempted indeed to see examples (24) and (25) as instances of compound forms, where the pronominal phrases appear as objects. But these rare instances may not be sufficient for drawing any conclusions on a more general basis, as the elements are often separated from each other by other sentence elements. Both each ... other and each one ... other constructions appear even in the first sub-period of the HC (eachone in (32) and (34)), but again, the instances are very few. And each ... other without one in examples (31), (33) and (34).

Visser's time scheme for reciprocal expressions provides interesting material for comparison (1970:439 A.A.9). According to him, each ... other and either (etc.) ... other make themselves apparent from Old English on, whereas the first instances of the type every (etc.) ... other are from c. 1290, and each ... the other from as late as 1586. He also gives the earliest date for the one ... the other as c. 1205. One ... another has only been recorded from 1548 onwards.

To cover the gap between Old and Middle English uses and the Present-Day English usage, reference should also be briefly made to a study by Helena Raumolin-Brunberg (1997), whose article on reciprocals focuses on how the two pronouns acquire their compound character, from discontinuity to unity, during the Early Modern English period. Interesting material can also be found in Sheen's work (1988). Raumolin-Brunberg points out some constraints that may have played a role in their development. Using the CEEC (Corpus of Early English Correspondence) and the EModE part of the HC she bas found 126 instances of reciprocals in the CEEC and 85 in the HC. The majority of the instances represent discontinuity, i.e. the examples clearly consist of two elements; in most cases the expression is divided between subject (note the difference to PDE usage) and object functions. But discontinuity is particularly well illustrated by expressions with a preposition, which normally comes between the two elements like in:

37) It is good to quicken these affections in these affections in these dead declining tymes, and when on one another in love and fayth which are the bonds of perfection uniting us one unto another, and all to Christ ... (CEEC, 1629, Sir William Masham 77)

where we have both types and in the following examples, out of which the last mentioned one also represents the genitive use of the reciprocal. (All examples quoted from Raumolino-Brunberg 1997).

38) A plague upon it when thieves cannot be true one to another. (Henry IV, Part II, ii,28)

39) ... thickets lined with most pleasant shade, in which the nightingales strive one with the other which should in most dainty variety excel one the other. (CEEC, 1640, Henry Oxinden 166)

40) and setche a lawse company that thay had worke annough to pyke of lyse oon oof onother's clothes. (CEEC, 1546, Sabine Johnson 788) Raumolin-Brunberg quotes an example of 1592 as the earliest instance of one another, but an approximately 20 years earlier instance has been quoted by Sheen (1988):

41) The shepherds there robben one another. (Spenser, Shepherd's Calendar, 38)

Both the CEEC and HC show growth in the general frequency of the reciprocal pronouns. According to Raumolin-Brunberg (1997: 231) of the three alternatives one another, (the) one the other and each other, one another is the most frequent until the last section of the HC (1640-1710), during which each other becomes slightly more common. (This provides an interesting case to compare with the OE material, which shows contrary evidence.) On the basis of the CEEC data it is impossible to say anything else about the significant factors contributing to this difference in frequencies than that one another seems to be the most colloquiat alternative (cf. Biber et al. (1999) who claim the contrary for PDE usage), and that individual preferences seem to be more relevant than genres or social groups. This again, seems to be contrary to the present-day usage in which, according to Biber et al., (1999) one another is the least colloquial reciprocal expression.

When do the compound forms each other and one another emerge and get established? One another begins to develop into one unit after 1620 (Raumolin-Brunberg 1997: 233), but variation still exists even after that. According to Sheen (1988: 271), one another appears in subject position in poetry in Spenser and Shakespeare, but then not as a compound (as one unit). One ... the other does hot become a compound at all. Each other, on the other hand, has practically completed the process into one unit by Restoration times. Malory, Caxton and Shakespeare use it even earlier, but again as a compound form only in object and not in subject position.

To conclude, it remains to be said that the fascination of English reciprocals lies in the fact that they have undergone such a radical change from OE to PDE and, contrary to the general tendency not to develop synonymous expressions, English has, to quite an extent, preserved a pair of such, and quite a complex system to express reciprocity in general. The system has in some ways been simplified from OE and some of the EModE variants have acquired specialized uses (like every other or either). But as we would not like to believe in free variation, we might indeed try to sec specialization here. Also, as noted earlier, stylistic differences seem to be relevant in ModE usage. We are also here dealing with a grammaticalization process in which grammaticalised items have become even more grammaticalised: separate pronominal elements have grown into a compound pronoun. I feel very tempted to sec an on-going subjectification process in PDE, if we tan indeed trace the subtle stylisfic shades, noted, e.g., by Biber et al. (1999). Unfortunately for the OE and ME periods such delicate distinctions are hard to detect because of the very limited material, which does hot tie neatly with such stylistic or social factors.

(1) Kjellmer illustrates this hierarchical structure of the rules graphically in his article (1982: 254).

(2) The data have been retrieved from The Helsinki Corpus of English Texts (HC) and the Toronto material for the Dictionary of Old English.

(3) The role of the accusative pronoun with an infinitive is a very complex matter and has been discussed thoroughly by Mitchell (1985: [subsection] 3734-40).

(4) Sheen's extensive and detailed lists include combinations like EACH/OTHER, EACH+OTHER, EACH+PREP+OTHER, EACH+PREP+THE OTHER, EACH OF (US, YOU, THEM)+OTHER, etc; EVERY+OTHER, EVERY MAN+OTHER, EVERY ONE+ V+OTHERS, EVERY ONE+PREP+OTH ER, etc.

REFERENCES

Biber, Douglas--Stig Johansson--Geoffrey Leech--Susan Conrad--Edward Finegan

1999 Longraan grammar of spoken and written English. London: Longraan.

Blake, Nomlan (ed.)

1999 The Cambridge history of the English language. Vol. 2. 1066-1476. Cambridge: Cambridge University Press.

Curme, George O.

1931 Syntax. Boston: Heath.

1935 Parts of speech and accidence. Boston: Heath.

Dalrymple, Mary--Makoto Kanazawa--Yookyung Kim--Sam Mehombo--Stanley Peters

1998 "Reciprocal expressions and the concept of reciprocity", Linguistics and Philosophy 21: 159-210.

Heine, Bernd--Ulrike Claudi--Friederike Hunnemeyer

1991 Grammaticalization. A conceptual framework. Chicago/London: University of Chicago Press.

Helsinki Corpus of English Texts

1991 Helsinki: Department of English, University of Helsinki.

Hogg, Richard M. (ed.)

1992 The Cambridge history of the English language. Vol. 1. The beginnings to 1066. Cambridge: Cambridge University Press.

Kahlas-Tarkka, Leena

1987 The uses and shades of meaning for every and each in Old English. (Memoires de la Societe Neophilologique de Helsinki 46). Helsinki: Societe Neophilologique.

Kjellmer, Goran

1982 "Each other and one another: On the use of the reciprocal pronouns in English", English Studies 63, 3: 231-254.

Kruisinga, E.--P.A. Erades

1953 An English grammar Groningen: P. Noordhoff.

Kyto, Merja

1996 Manual to the diachronic part of the Helsinki Corpus of English Texts: Coding conventions and lists of source texts. (3rd edition). Helsinki: Department of Eng[ish, University of Helsinki.

Lass, Roger (ed.)

1999 The Cambridge history of the English language. Vol. 3. 1476-1776. Cambridge: Cambridge University Press.

Mitchell, Bruce

1985 Old English syntax. Oxford: Clarendon.

Mustanoja, Tauno F.

1960 A Middle English syntax. Part I. Parts of speech. (Memoires de la Societe Neophilologique de Helsinki 23). Helsinki: Societe Neophilologique.

Porter, S.

1953 "The expression of reciprocity", English Studios 34: 252-257.

Poutsma, H.

1916 A grammar of Late Modern English. Groningen: P. Noordhoff.

Quirk, Randolph et al.

1985 A comprehensive grammar of the English language. London: Longman.

Raumolin-Brunberg, Helena

1997 "Reciprocal pronouns: From discontinuity to unity", Studia Anglica Posnaniensia XXXI: 227-236.

Raumolin-Brunberg, Helena--Leena Kahlas-Tarkka

1997 "Indefinite pronouns with singular human reference", in: Matti Rissanen--Merja Kyto--Kirsi Heikkonen (eds.), 17-85.

Rissanen, Matti

1967 The uses of one in Old and Early Middle English. (Memoires de la Societe Neophilologique de Helsinki 31). Helsinki: Societe Neophilologique.

Rissanen, Matti--Merja Kyto--Kirsi Heikkonen (eds.)

1997 Grammaticalization at work. Berlin--New York: Mouton de Gruyter.

Schrader, B.

1887 Studien zut AEfricschen Syntax. Diss. Gottingen.

Sheen, Ding-Taou

1988 The historical development of reciprocal pronouns in Middle English with selected Early Modern English comparisons. [Unpublished PhD thesis, Ball State University at Muncie, Indiana].

Visser, E Th.

1963-1973 An historical syntax of the English Language. Vols. I-III. Leiden: Brill.

LEENA KAHLAS-TARKKA

University of Helsinki