A focus on adverbial connectors: Connecting, partitioning and focusing attention in the history of English [1]

Ursula Lenker, Catholic University of Eichstätt-Ingolstadt

1. Introduction

In his seminal and exhaustive grammar of Old English, Bruce Mitchell cites Quirk/Wrenn’s Grammar stating that ‘the free variation available to Ælfric in the position of the adverbs is available today likewise’ and adds ‘So much depends on the writer’s purpose’ (Quirk & Wrenn 1977: 91; Mitchell 1985: §§ 1592–1593). Both of these findings, i.e., that different, though not the same, options of placement of adverbials have been available since the Old English period and, more importantly, that writers and speakers employ different positions of adverbials for different purposes are not as trivial as they may appear to be at first glance. The history of English has seen substantial changes in word order, resulting in the fact that in Present-Day English different patterns of word order may only very restrictedly be employed for different ‘purposes’: Modern English word order is almost invariably fixed to SVO. This is in striking contrast to Old English and Middle English and also to the present-day versions of the other Germanic languages, which allow speakers and writers to signal particular kinds of information structure by, for example, a fronted object (OVS). In Present-Day English, by contrast, the only remaining flexible sentence constituent is the adverbial: It is the only constituent which can be placed initially, medially and finally, though different positions are favoured for different kinds of adverbials, i.e., for time, place or manner adjuncts and, in particular, for disjuncts and conjuncts (see section 4.1).

Mitchell does not specify the ‘purposes’ of the Old English writers with regard to adverbial placement in any detail. Generally, there has – at least until recently – been little research into the placement of adverbials in the history of English, since – at least for the early stages of English – the position of the verb (V-final, V2) and, relatedly, the object (VO, OV) and the fixing of the word order to SVO and its consequences have been in the centre of diachronic word order studies in English. In some of these studies, adverbial position – in particular the initial position – is given some attention when triggering inversion, i.e., AVS/O (see, for example, chapters,, and 6.1.1 in Swan 1988 or Bækken 1998: 186–295).

As regards the discourse-structural functions of Present-Day English adverbials, research has until recently predominantly dealt with different functions of pre- vs. post-posed adverbial clauses (see, for example, Thompson 1985, Ford 1993, Diessel 2005). The only comprehensive study of one class of adverbials – namely adjuncts – is Hasselgård (2010), who investigates different positions of all realizations of adjuncts in a corpus-based study and, as one of the first linguists, also discusses their functions with respect to thematic organization and information structure (see in particular Hasselgård 2010: 73–95, 102–114, 123–151).

In cross-linguistic as well as theoretical accounts of word order, linguists have lately started to investigate the diagnostic value of adverbial placement in general (see the summary in Svenonius 2002: 201–209). As concerns the diachrony of English, the most important approach for our investigation is the one by van Kemenade, Los and collaborators (see, for instance, van Kemenade & Los 2006 or van Kemenade, Milicev & Baayen 2008) suggesting that some Old English adverbs – þa and þonne ‘then’ in particular – should be analysed as ‘discourse partitioners’ or focus markers – i.e., markers distinguishing topic from focus material. [2]

In the present chapter, I offer a preliminary investigation into the changes in the placement and word order patterns of one specific kind of adverbs, adverbial connectors (Biber et al. 1999 ‘linking adverbials’, Quirk et al. 1985 ‘conjuncts’). I will first summarize the most important findings for the history of adverbial connectors in English (based on Lenker 2010) and will then show how these changes have affected the possibilities of positional variation of these adverbs, focusing on changes since the end of the Early Modern English period. In the concluding section, I will suggest that the changes in form, morphological make-up and semantics of these adverbials until the Early Modern English period allow their employment not only for explicitly marking cohesion, but for two different ways of marking information structure in written texts. In certain medial positions (e.g., after the copula or after the first auxiliary), the adverbial connector – in addition to its connecting function – partitions GIVEN and NEW material. Secondly, another medial position, the ‘post-initial position’ will be singled out: This position of the adverbial connector (usually following the subject in Modern English) – in addition to its connecting function – serves to focus attention on the first element in the sentence, similar in function to focus adverbs such as only, particularly, etc. (see section 4.3.1).

2. Adverbial connectors in the history of English

2.1 Connectors: Definitions – types of connectors

In the present study, connectors are defined as linguistic items which signal a two-place relation between segments of text above the level of the phrase, i.e., between sentences or chunks of discourse, here called ‘connects’ (for this term see Pasch et al. 2003). The meaning of a connector is procedural, not conceptual: it does not change the propositional content of any of the segments it relates. Connectors thus reveal or make explicit the connections already operating in a text.

A differentiation of three different kinds of connectors is basically agreed on in the literature: ‘coordinating conjunctions’ (working on the clausal/textual level, not on the phrasal level), ‘subordinating conjunctions’ (‘subordinators’) and ‘adverbial connectors’ (Quirk et al. 1985 ‘conjuncts’, Biber et al. 1999 ‘linking adverbials’). The – mainly topological – criteria traditionally employed for distinguishing the different types of connectors are (i) the position of the connector, (ii) the sequence and position of the respective connects, (iii) the possibility of collocations of connectors, and (iv), in languages such as German, constituent order (V2 in main clauses and V-final in subordinate clauses; for details, see Lenker 2010: 22–32).

Table 1. contrastive/concessive connectors in Present-Day English.

(Syndetic) Parataxis

Coordinating Conjunction:

(1) He tried hard, but he failed.

Adverbial Connector:

(2) He tried hard, and yet he failed. [and + Adverbial Connector]

Adverbial Connector:

(3a) He tried hard. However, he failed. [initial]

(3b) He tried hard. This time, however, he failed.


(3c) He tried hard; he failed, however.



Subordinating Conjunction:

(4a) (Al)though he tried hard, he failed. [preposed]

(4b) He failed, though he had tried hard.


While coordinating and subordinating conjunctions are only found clause-initially (see Present-Day English but, and, although and though in examples (1), (2) and (4a-b)), adverbial connectors are, like most kinds of adverbs, more flexible in their position (see however in examples (3a-c); for restrictions on certain adverbials such as yet, see section 4.2.3). As concerns the order of connects (criterion (ii)), parataxis usually requires a fixed order of connects. Accordingly, clauses linked by coordinating conjunctions and adverbial connectors only allow one sequence of connects (see examples (1), (2) and (3)). Subordinate clauses, on the other hand, may be placed before or after the superordinate clause (see English (al)though in examples (4a) and (4b)). Coordinating conjunctions such as and or but may – in contrast to the other types of connectors (criterion (iii)) – collocate with subordinators and also with adverbial connectors (see English and yet in example (2)).

2.2 The history of adverbial connectors in English: Summary of findings

While the inventory of the coordinating conjunctions relevant for the present study, i.e., those working on the sentence or discourse level, has remained fairly stable throughout the history of English (Present-Day English and, but and or), [3] there have been drastic changes in the categories of both subordinators and adverbial connectors.

In Old English, most of the connectors are polyfunctional: some of them are circumstance adverbs (time/space adverbs) such as nu ‘now’, þa ‘then, there’ or þonne ‘then’ (cf. the similar polyfunctionality of Present-Day English now and then) or epistemic adverbs such as eornostlice ‘earnestly’, soþlice/witodlice ‘truly’ with a context-dependent connector force (see Lenker 2010: 58–59). The greater part of connectors, in particular those used most frequently, are so-called ‘ambiguous adverbs/conjunctions’ (for the term see Mitchell 1985: § 2536). These ‘ambiguous adverbs/conjunctions’ may be used as adverbial connectors or subordinators (for a closer analysis and examples see Lenker 2010: 58–67):


Adverb ‘also, moreover’ addition

Conjunction ‘as, so’

eac (swylce)/swylce eac

Adverb ‘also, moreover’ addition

Conjunction ‘all such …, as’

forþæm (forþon, forþy)

Adverb ‘for; therefore’ cause/result

Conjunction ‘because’


Adverb ‘now’ transition

Conjunction ‘now that …’


Adverb ‘so’ cause/result

Conjunction ‘so that …; as …’


Adverb ‘then’ transition

Conjunction ‘then … when’


Adverb ‘then’ transition, cause

Conjunction ‘then … when’

One of the most common Old English connectors, causal/resultive forþæm/forþon/forþy, for example, may be used as a resultive adverbial connector (cf. Present-Day English therefore, consequently), as in


þas þing we gemetton on Ramesige þurh Godes miltsigendan gife. Forþan ic ne swigie for ðæra bocre getingnyssum ne for þæra gelæredra manna þingum þe …

‘We found these things at Ramsey through God’s merciful grace. Therefore I shall not be silent either on account of the eloquence of the literate or for the sake of those learned men who …’ (ByrM 1.1.158)

It can, however, also be used as a causal subordinating conjunction (cf. Present-Day English because, since or as), as in


Ðas þing we swa hwonlice her hrepiað on foreweardum worce forðan we hig þenceað oftor to hrepian and to gemunanne.

‘We discuss these things so briefly at the beginning of this work because we intend to discuss and recall them more often.’ (ByrM 1.2.250)

Since the distinction of adverbial vs. conjunction is not regularly indicated by different word order patterns (i.e., a distinction of V2 for the main clause vs. V-final for the subordinate clause) or morphological features, these items are indeed ambiguous (for details and certain correlations such as the more frequent use of the particle þe, such as in forþæm þe, in subordinate clauses, see Lenker 2010: 140–165).

Because of the great number of Old English ambiguous adverbs/conjunctions, the general line of development is initially to some extent parallel for adverbial connectors and subordinators: In the course of an increased form-to-function mapping after the Old English period, most of the Old English ‘ambiguous adverbs/conjunctions’ are discarded, so that we see a corresponding decrease in syntactic polyfunctionality and also in semantic polyfunctionality in both of the – now separate – classes of adverbial connectors and subordinators.

As concerns their later history, there is a stark difference in the survival of the innovations in the class of subordinators on the one hand and adverbial connectors on the other: The inventory for subordinators was – in addition to the repertoire of coordinating conjunctions (see endnote 3) – set to almost three quarters (ca. 73 per cent) by the end of the Middle English period (see Kortmann 1997: 291–335). For the relations cause/result, for example, Kortmann finds that ‘[b]y and large, then, the inventory of causal connectors in Present-Day English resembles that of (Late) Middle English most closely’ (1997: 331). The findings are strikingly different for adverbial connectors: more than sixty per cent of today’s adverbial connectors were only coined in the (late) Early Modern, Late Modern and Present-Day English period. In contrast to subordinators, for which Middle English was the decisive period, there is no single period which can be established as crucial for the Present-Day English inventory of adverbial connectors. A more detailed investigation of new coinages and losses, however, shows that with respect to adverbial connectors Middle English emerges as a ‘period of experiment and transition’, i.e., most of the adverbial connectors coined in Middle English have not survived. This is mainly due to the collapse of the system of Old English ‘pronominal connectors’, i.e., connectors comprising a case form of the demonstrative (e.g., the dative þæm) which became opaque in Middle English (cf., for example, forms such as Middle English forthen; see MED, s.v. for-than; for details see Lenker 2010: 96–102, 154–167). Only from the end of the Early Modern English period onwards has the group of adverbial connectors been relatively stable (see Lenker 2010: 77–87).

Table 2. Adverbial connectors in the history of English. ° = ambiguous adverb/conjunction. [4]

Survivors to PDE

Lost on their way to PDE


°eall-swa ‘also’, elles ‘else’, her ‘here’, °nu ‘now’, °swa ‘so’, °þeah ‘though’, °þonne ‘then’

ærest ‘first’, eac ‘also’, °eac swylce/swylce eac ‘also’, cuþlice ‘certainly’ > ‘therefore’, eft ‘also’, eftsona/eftsones ‘also’, eornostlice ‘earnestly’ > transition, °forþæm/°forþon ‘therefore’, °forþy ‘therefore’, huru ‘nevertheless’, huruþinga ‘nevertheless’, naþylæs ‘nevertheless’, °hwæt þa ‘what then’ > transition, °hwæðere ‘nevertheless’, soþlice ‘truly’ > transition, swaþeah ‘nevertheless’, swaþeahhwæðre ‘nevertheless’, °þa ‘then’ > transition, þa giet ‘also’, (þær)toeacan ‘also’, þæs Demonstrative Genitive > ‘therefore’, þeahhwæðere / hwæðreþeah ‘nevertheless’, °þy Demonstrative Instrumental > ‘therefore’, gewislice ‘certainly’ > transition, witodlice ‘truly’ > transition

ME 1/2

certainly, first, °for, at (the) last, more, overall, therefore, °yet

certain, certes, forsooth, herefore, iwis, for that, °forhwi, neverthelatter, nought-for-than ‘nevertheless’, on oðer half, oðer side, °wherefore

ME 3

further, farthermore/furthermore, moreover, nevertheless, otherwise, thus

°algates, ergo, for which, fartherover/furtherover, item, °nought-for-that ‘nevertheless’, °nought-for-thi ‘nevertheless’, overmore, over this/that, otherways, sekirly, suingly ‘consequently’, truly, verily

ME 4

contrariwise, finally

after ‘moreover’, °howbeit, notwithstanding


again, consequently, contrarily, likewise, nonetheless, surely, well

°albeit, in fine, °howsomever, likeways, over and besides, semblably


above all, besides, however, indeed, at least, (in the) meantime, (in the) meanwhile, in sum, sure

beside, contrary, contrariways, °howsoever, lastly, over and above


after all, on the other hand, hence, still

thence, videlicet, whence


accordingly, at the same time, too

in truth


altogether, anyhow, conversely, of course, at all events, in fact


all the same, in any case, anyway, at any rate, in the same way


all in all, correspondingly, incidentally, instead, plus, rather, similarly

This comparison of inventories of adverbial connectors through the history of English shows that most of the adverbial connectors of today’s English were only coined in the Early Modern English period or later. As concerns their morphological make-up, most of them are – in the terminology of the Longman Grammar of Spoken and Written English (LGSWE) – ‘single adverbs’ (Biber et al. 1999). The LGSWE’s quantitative studies find that, with regard to their syntactic realization, linking adverbials ‘display the strongest association with a single syntactic form: almost 80 per cent of the linking adverbials are realized by single adverbs’ (Biber et al. 1999: 768). A closer examination of the items shows, however, that new adverbial connectors coined after the abandonment of ambiguous adverbs/conjunctions and the collapse of the system of pronominal connectors are only rarely simple transfers of purely lexical adverbs (e.g., circumstance adverbs) to linking adverbials (see the very few cases of items such as accordingly, consequently or conversely). Most of the new adverbial connectors are lexicalized phrases such as altogether or all the same or, in particular, lexicalized prepositional phrases such as indeed, instead or after all (for details, see Lenker 2010: 94–98). These phrases typically comprise material from two cognitive source domains – time and space. See, for example, the following survey of surviving adverbial connectors coined from Early Modern English onwards (for details, see Lenker 2010: 106–113):

time EModE after all, however, (in the) meantime, meanwhile
LModE at the same time
place/space EModE above all, again (cf. OE on-gean, PDE against), besides, hence, over and above
LModE anyway, in the same way
PDE instead

These two domains time and space have been used as source domains for adverbial connectors since the Old English period, in particular in a connector use of circumstance adverbs such as temporal OE nu ‘now’ or OE þonne ‘then’, or spatial her ‘here’ or temporal/spatial OE þa ‘then; there’ (see Lenker 2010: 108–113). Both of these domains are very useful for connectors as markers of textual cohesion, since they provide the necessary discourse-deictic references for the explicit linkage of sentences or chunks of discourse – a function formerly fulfilled by inflected pronouns (cf. dative demonstrative þæm in OE forþæm) in the pronominal connectors (see section 2.2).

As a result of their morphological history, most of the adverbial connectors of Present-Day English are unambiguous in their adverbial connector function (the only exceptions are some circumstance adverbs such as now, then, accordingly, etc.). Their monofunctionality allows them to be placed at different positions in the sentence, usually without change in meaning or scope (in contrast to, for example, adjuncts/disjuncts such as frankly – cf. He told me frankly vs. Frankly, he told me). Adverbial connectors are thus a particularly apt means for fulfilling further discourse functions in addition to their basic connecting function, namely the functions of focusing attention on particular sentence elements or their use as partitioners.

3. Positional flexibility of adverbs and textual organization in Present-Day English

3.1 Introductory remarks

At the beginning of this study, it was pointed out that the changes in the adverbial connectors have not only affected their semantics (cf. the loss of polyfunctionality) and morphological make-up, but also their positional flexibility. This becomes particularly manifest from the middle of the eighteenth century onwards. I will argue here for the first time that this date is by no means a coincidence, but that these relatively recent changes in the placement of adverbial connectors are related to the fixation of word order to SVO and the subsequent decreasing possibilities of indicating different thematic structures by different word orders. This increased positional flexibility in the placement of adverbial connectors, which affects the written mode only, will here be suggested to serve as a new means for indicating thematic structure and textual organization in English after the loss of V2 and the fixation of word order to SVO. Different medial positions of adverbial connectors will be shown to serve as distinct signals of information packaging and thus to be similar in function to other fairly recent information packaging structures, such as new cleft-sentence or passive constructions or a wider flexibility in the kinds of subjects (see, for example, Los 2009 and the research project ‘Syntax and Information Structure: Discourse Options after the Loss of Verb-Second (NWO), 2008-2013’ (Los & van Kemenade)).

I will here concentrate on the changes after the Middle English period. The situation in Old English and early Middle English is much more complicated, precisely because of the word order flexibility in these periods and also because of the prominence and special character of OE þa (‘then, when; there’ or transition connective), the shibboleth of Old English narrative style (see Lenker 2010: 64–66 and 238–239 and also Wårvik, this volume).

3.2 Collocations (initial position) vs. medial position of adverbial connectors

Let me introduce the basic changes in the placement of adverbial connectors by sample passages from texts of a comparable text type, i.e. treatises, by renowned, mature authors of their respective periods, whom we would normally not accuse of ‘not knowing how to write better prose’. (7) is taken from the Tale of Melibee by Geoffrey Chaucer (ME3; 1343?–1400), and (8) is from Adam Smith’s The Wealth of Nations (LModE1; 1723–1790).


but certes (7a) what ende that shal therof bifalle, it is nat light to knowe. For soothly (7b), whan that werre is ones bigonne, ther is ful many a child unborn of his mooder that shal sterve yong by cause of thilke werre, or elles lyve in sorwe and dye in wrecchednesse. And therefore (7c), er that any werre bigynne, men moste have greet conseil and greet deliberacion." And whan this olde man wende to enforcen his tale by resons, wel ny alle atones bigonne they to rise for to breken his tale, and beden hym ful ofte his wordes for to abregge. For soothly (7d), he that precheth to hem that listen nat heeren his wordes, his sermon hem anoieth. (CMCTPROS, Chaucer, Tale of Melibee, p. 219.C2).


This portion, however (8a), may still be considered as the natural rent of land, or the rent at which it is naturally meant that land should, for the most part, be let. The rent of land, it may be thought (8b), is frequently no more than a reasonable profit or interest for the stock laid out by the landlord upon its improvement. This, no doubt (8c), may be partly the case upon some occasions; for (8d) it can scarce ever be more than partly the case. The landlord demands a rent even for unimproved land, and the supposed interest or profit upon the expense of improvement is generally an addition to this original rent. Those improvements, besides (8e), are not always made by the stock of the landlord, but sometimes by that of the tenant. When the lease comes to be renewed, (8f) however, the landlord commonly demands the same augmentation of rent …

Hence (8g) a greater rent becomes due to the landlord. It requires, too (8h), a more attentive and skilful management. Hence (8i) a greater profit becomes due to the farmer. The crop, too (8j), at least in the hop and fruit garden, is more precarious. Its price, therefore (8k), besides compensating all occasional losses, must afford something like the profit of insurance. (CLSMI1, The Wealth of Nations, Chapter XI, Part 1).

The differences between these two passages – both of which are prototypical of their respective period – are obvious at first glance: in the passage from Chaucer’s Tale of Melibee, all of the sentences start with explicit markers of textual cohesion. In all cases, these explicit markers are not just conjunctions or adverbial connectors, but collocations of a conjunction (which functions as a rather loose connective) and an adverbial connector. The adverbial connector may express an additional semantic relation such as cause (and therefore in (7c)) or transition (but certes in (7a) or for soothly in (7b) and (7d)). Collocations like these are typical of Chaucer’s prose (Kerkhof 1982: 456 and 459). Unfortunately, with the exception of the introductory passages of the Treatise on the Astrolabe (CMASTRO), all of Chaucer’s prose works are translations from Latin or French (Boece, Tale of Melibee; Parson’s Tale), so that we cannot rule out loan influence in these collocations. Yet, the predilection for collocations is also widely attested in prose texts which were originally composed in Middle English, such as Caxton’s Prologues:


For (9a) in the sayd boke they may see what this transitorie & mutable worlde is And wherto euery mann liuyng in hit / ought to entende Thenne for as moche as (9b) this sayd boke so translated is rare & not spred ne knowen as it is digne and worthy For the erudicion and lernyng of suche as ben Ignoraunt & not knowyng of it … And furthermore (9c) I desire & require you … and therfore (9d) he ought eternelly to be remembrid. of whom the body and corps lieth buried … (CMCAXPRO, p. 63).

By contrast, not a single one of these collocations is found in the passages from Adam Smith’s The Wealth of Nations (1776). While we also find sentence-initial connectors, such as causal/transitional for (8d) and two instances of resultive hence (8g, 8i) there, all of the other adverbial connectors are placed sentence-medially, such as reinforcing besides (8e) and too (8h, 8j), contrastive however (8a, 8f), and resultive therefore (8k). Most of these are placed in the position after the first constituent (after the subject in (8a), (8b), (8e), (8j), (8k) and after an adverbial in (8f)), a position which will below be referred to as the post-initial-position (see section 4.2.3); only in (8h), additive too is placed between the verb and the object.

3.3 Placement of adverbial connectors from Old English to Late Modern English

The examples above seem to indicate that the medial position of adverbial connectors has only been attested in higher frequencies after the Middle English period. This assertion was put to the test in a corpus study, using one of the corpora compiled for the diachronic studies of adverbial connectors in Lenker (2010), viz., a balanced corpus of “treatises” and “sermons/homilies” (for details, see Lenker 2010, Appendix C.2). This corpus consists of samples of either four or five texts, each comprising ca. 5,000 words (i.e., altogether ca. 20,000 to 25,000 words per sub-period). If at all possible, complete texts were chosen.

The basic selection of texts for both of the corpora used in Lenker (2010) was based on quantitative findings for Present-Day English. The investigations in the LGSWE (Biber et al. 1999: 765–776 and 880–892) show that – as concerns the LGSWE’s core registers conversation, fiction, newspaper language and academic prose – linking adverbials are most common in academic prose (for corresponding findings on a smaller data set, see Greenbaum 1969: 79–80). The relatively high frequency of adverbial connectors in academic prose is, of course, not a coincidence. The main communicative purpose of these texts is information, argumentation and explanation for a specialist audience, i.e., the emphasis is put on conveying logical and unambiguous coherence. The text types chosen for the diachronic corpus were thus those whose situational properties are most similar to LGSWE’s category academic prose: academic or scientific language, i.e., homilies or religious, philosophical, educational and literary treatises (on the exclusion of poetry and narratives and test analyses of the use of adverbial connectors in these genres see Lenker 2010: 12–16).

As a first step, the corpus analysis focuses on the quantitative relation of the coordinators and, but (OE ac) and for (all of which – as coordinating conjunctions – obligatorily have to be placed sentence-initially) on the one hand and adverbial connectors on the other hand (on the exclusion of OE þa, see Lenker 2010: 64–66, 238–239 and section 3.1).

Figure 1. Mean frequencies (/ 10,000 words) of conjunctions vs. adverbial connectors (without OE þa) in the corpus of “treatises” and “sermons/homilies” (Lenker 2010: Appendix C).

After a sharp rise after the Old English period, we see a relatively steady decrease in the number of conjunctions from Middle English to the late Early Modern English period, followed by a sharp decrease (to about half of the instances of LMod2) in Late Modern English 3, i.e., the English of the nineteenth century (1850–1920). This is mainly due to the decreasing frequencies of and (see Culpeper & Kytö 2000 and Meurman-Solin, this volume) and ‘recursive for’ (for a discussion of recursive for as a loose connective in late Middle and Early Modern English, see Lenker 2010: 161–164 and Meurman-Solin forthcoming).

If we have a closer look at the positions of the connectors, however, we notice that the differences between Middle English and Late Modern English are even more pronounced. Furthermore, they are not only due to an increasing use of a wider range of different adverbial connectors, but to deliberate choices of authors not to use sentence-initial collocations consisting of a semantically looser connective and a semantically more explicit connector, but medial adverbial connectors. Figure 2 shows that the impression we have gained from examples (7) to (9) above is indeed supported by the quantitative analysis of the corpus texts.

Figure 2. Mean frequencies (/ 10,000 words) of collocations (conjunction + adverb in initial position) vs. adverbial connectors in medial position (without OE þa) in the corpus of “treatises” and “sermons/homilies” (Lenker 2010: Appendix C).

In Figure 2, we see two periods emerge as decisive. First, the beginning of Middle English saw a rapid increase in the number of sentences which are introduced by a conjunction (and, but or for) in a collocation together with an adverbial connector (see example 7 above). This again attests to the difference of connector use in Old English in contrast to the other periods, even if we exclude OE þa. ME3 (1350–1420) shows the highest frequencies of such collocations. This process can be explained by the attempts at the evolution of a new English prose style in the genres of “treatises” and “sermons/homilies”, which eventually replaced the specific Old English prose style during the fourteenth and fifteenth centuries, i.e., only after the time when English had not been used in the written medium – in particular in official registers and prose literature – for almost two centuries, but had been supplanted by Latin and French (see Mueller 1984 and Meurman-Solin, this volume; on connective profiles of “treatises” and “sermons/homilies” see Kohnen 2007).

The second decisive period appears in the language of the late eighteenth and early nineteenth century (LModE2 from 1780–1850), when we see a reversal of the distribution of sentence-initial collocations and medial adverbial connectors. The difference is not that initial conjunctions or adverbial connectors are no longer used in Present-Day English academic prose, but that the medial placement of adverbial connectors becomes frequent. This is also confirmed by corpus findings for the position of adverbial connectors in Present-Day English:

% in initial position % in medial position % in final position
CONV ■■■■■■■■■■■ ■■■■■■■■
ACAD ■■■■■■■■■■ ■■■■■■■■ ■■

Figure 3. Positions of linking adverbials in conversation and academic prose according to Biber et al. (1999: 889, Table 10.18): ■ = 5 %; □ = less than 2.5 %).

In the corpus of the Longman Grammar, medial adverbial connectors are predominantly, indeed almost exclusively, attested for the written genres; in less than 2.5 per cent of all instances are adverbial connectors placed sentence-medially in conversation (for corresponding findings on a smaller data set, see Greenbaum 1969: 79–80). Academic prose clearly favours the initial (ca. 50 per cent) and medial positions (more than 40 per cent). The initial position can thus still be considered the unmarked position for adverbial connectors in this genre (see Greenbaum 1969: 80 and Biber et al. 1999: 891). Yet, not very much less, viz. more than 40 per cent of the adverbial connectors are placed in medial position. The medial positioning of adverbial connectors thus has become one of the most specific properties of the English academic prose style in the nineteenth and twentieth centuries.

In my earlier study I have tried to relate these findings to the changing cultural climate at the end of the Early Modern English period triggered by the idea of perspicuitas, which replaces copia as the leading style paradigm in the eighteenth century (see Lenker 2010: 241–246). Adverbial connectors, which make the speaker’s commitment as to the relations of discourse segments explicit, are obviously a very apt means of fostering perspicuity in language. The spirit of perspicuity also led to a more systematic punctuation, which in turn allowed or at least facilitated the unambiguous medial positioning of adverbial connectors in the written mode (see Parkes 1992: 89–91). The commas, which commonly signal the status of an adverb as an adverbial connector (and not, if there should be one of the rare cases of ambiguity, a circumstance adverbial), are seen to reflect pauses in the same construction in the spoken language in most of the literature (see Lenker 2010: 281–282). This, however, can only be considered a case of orthographic analogy (i.e., commas are the punctuation marks which, as a rule, signal pauses), since the medial positioning of connectors is only very rarely attested in spoken language. It might thus rather be argued that these commas should be regarded as signs underscoring the discourse structure of the second connect, marking either ‘a focus of attention’ on the preceding element or a ‘partition’ of given and new material, a function fulfilled by prosodic features (stress, jump in pitch) in the spoken mode (see Greenbaum 1969: 195 and section 4.2.2).

While I still think that the promotion of the new positions of adverbial connectors is related to these culturally inspired rhetorical ideas and, in particular, the fixation of punctuation rules and the recommendations in George Campbell’s highly influential The Philosophy of Rhetoric ([1776] 1963), I would now like to argue that there are also language-internal factors which have led to this change in adverbial placement in the written mode in the eighteenth and nineteenth centuries. I would like to suggest that the changes may be viewed as a reaction to the – at that time basically accomplished – fixation of the English word order to SVO, leaving adverbials as the only flexible sentence constituent in Modern English in unmarked contexts. For a better understanding of the possibilities of medial placement of adverbials, research on Modern English will now be summarized.

4. Adverbial placement and information structure

4.1 Positions of adverbials in Present-Day English

In Present-Day English, adverbials are the only sentence constituents which are comparatively free as to their position in the sentence. Also in contrast to the other sentence constituents, adverbials are optional and more than one adverbial can occur in a sentence. While subject, verb and object(s) are sequentially fixed in Present-Day English, adverbials can take three major positions: initial (before the subject), medial (after the subject but before any object/predicative; for details, see section 4.3.1) and final.

The corpus studies of the LGSWE show that all of the three positions are common. There are, however, strong preferences for different kinds of adverbials, i.e., for adjuncts (Biber et al. 1999 ‘circumstance adverbials’), content/style disjuncts (Biber et al. 1999 ‘stance adverbials’) and conjuncts (Biber et al. 1999 ‘linking adverbials’). [5] Each class of adverbial has a strong preference for a particular position: the most frequently attested position for adjuncts is the final position (see also Hasselgård 2010: 291). Disjuncts are most commonly found in medial position. Conjuncts or linking adverbials favour – in all modes – initial position (Biber et al. 1999: 770–774).

These basic distributional preferences are obviously related to the different meanings, functions and, thus, scopes of these adverbials. Circumstance adverbials commonly have a scope over the phrase, often completing the meaning of the verb and must thus follow the verb (and therefore, in Present-Day English, also the subject). Only when circumstance adverbs (of frequency, manner, place or time, etc.) need to be stressed or when they are employed for scene-setting functions (see section 4.2.1) can they also be placed in initial position. Furthermore, more than two adverbials in end-position are usually avoided – in this case, one of them is usually placed in front or medial position. Stance adverbials, which typically have an extended scope over the proposition of the entire clause, may be placed rather freely. Linking adverbials, as we have seen above, are most often used in initial position, so that the connection between two connects is clearly signalled as the reader or hearer moves from the first to the second connect. [6] Another factor which has to be taken into account is the weight/length of the adverbial itself (in particular when realized by a clause) or of one of the other sentence constituents: Adjuncts, for example, may be found in medial instead of end position because of a long and/or heavy argument of the verb, typically a direct object (cf. Hasselgård 2010: 290; on these factors, see also Quirk et al. 1985: 492–493).

4.2 Discourse features of adverbials

4.2.1 Earlier research

Starting with Thompson’s analysis of initial versus final purpose clauses (Thompson 1985), there has been an increasing and continuing interest in the discourse factors determining the position of subordinate clauses (see, for instance, Haiman & Thompson 1988, Ford 1993 and Diessel 2005, who argues that both processing and information structure are relevant to constituent order). There is wide agreement in the literature that initial adverbial clauses state a problem within the context of expectations raised by the preceding discourse, to which the succeeding material (often many clauses) provides a solution (cf. To cool, place the loaf on a wire rack). Final purpose clauses, on the other hand, play a much more local role of stating the purpose for the action named in the immediately preceding clause (cf. Place the loaf on a wire rack to cool).

While this topic-forming capacity of pre-posed adverbial clauses is now commonly agreed on (see, for example, Meurman-Solin & Pahta 2006 and Meurman-Solin forthcoming) and the discussion in Lenker 2010: 28–34), there has only been some scattered research into the information-structural functions of the placement of non-clausal realizations of adverbials, viz. adverbs or nominal or prepositional phrases (see, for instance, Jacobson 1964, Horová 1976, Taglicht 1984 and Ungerer 1988). This paucity of studies is most probably due to the heterogeneity of the word class of adverbs and the varieties of functions they can fulfil on the phrase and clause level (for surveys, see Hasselgård 2010: 14–39 and Lenker 2010: 33–57) and, for practical reasons, to the high number of adverbials in all kinds of texts: Hasselgård (2010: 6), for example, counts 110,970 adverbials distributed over about 46,000 sentences in the one-million words of the British Component of the International Corpus of English (ICE-GB). Both of these factors make corpus studies of adverbials extremely difficult.

As an introductory illustration, let me exemplify some of the different information-structural functions of the placement of adverbials by the examples in Torremolinos and however in (10) and (11) (taken from Ungerer et al. 1984: 10–11):


Spanish food is different from English food. In Torremolinos, however, some restaurants serve fish and chips.

When the place adverbial in Torremolinos is placed initially as in (10), it serves a ‘scene-setting effect’ (Ungerer et al. 1984: 10) and indicates that the proposition in the rest of the sentence pertains particularly or exclusively to this place (cf. the topic-forming capacity of pre-posed adverbial clauses). In (10), there is a definite contrast, i.e., a restriction on the validity of the proposition of the preceding sentence, which is here emphasized by the linking adverbial however following the adjunct in Torremolinos (placed before the subject).

In (11), on the other hand, where the adjunct is placed finally, the focus of the statement is on the place adverbial itself. Here, the place adverbial conveys new information.


You needn’t do without your fish and chips completely. You can get them in Torremolinos.

In the most detailed corpus study of all types of realizations of one type of adverbials, namely adjuncts (cf. in Torremolinos), in Present-Day English, Hasselgård finds – in agreement with Biber et al. 1999 – that the ‘end position can be considered the default position for most semantic types of adjuncts’ (Hasselgård 2010: 115–151 and 290). In line with the default information structure, viz. given before new information, end position is used for adjuncts that constitute the culmination point of the action or the informational peak of the clause. Initial adjuncts (cf. (10)), by contrast, are found to be used to ‘set up an interpretational framework for the rest of the clause’ in that they give an interpretative background for the message or place a restriction on its truth or validity (cf. Hasselgård 2010: 67–95).

4.2.2 Medial position of adverbial connectors

In Present-Day English SVO, medial or mid-position usually refers to various positions before the main verb, i.e., between (i) the subject and the (first part of the) verb phrase, (ii) after an auxiliary, but before the main verb or (iii) between the verb phrase or some other obligatory element. [7] As concerns adverbial connectors, we usually only find little substantial information on the different positions of adverbial connectors. The Longman Grammar, for example, only states that ‘the common linking adverbials in acad [academic prose] – therefore, thus and however tend to occur in medial positions (when not in initial position). In particular, these forms often occur immediately following the subject’. LGSWE then states that ‘these forms also occur in other medial positions’ and lists as positions: ‘immediately following an operator’ and ‘between a verb and a complement clause’ (Biber et al. 1999: 892).

Such a list of positional variants without specifying any reasons for the different positions and their functions arouses suspicion: The main function of adverbial connectors is that they – explicitly and unambiguously – signal the connection between two sentences or chunks of discourse. Information processing would thus be much easier if speakers chose to stick to one and the same position, most likely the initial position so that the relations – in particular relations of contrast and concession – are clearly signalled at the beginning of the second connect. For that reason the initial position is, as we have seen above, also the most common and unmarked position for adverbial connectors. Furthermore, in the online-production and processing of speech in the spoken mode, speakers indeed stick to one position, the initial one, with the exception of some adverbials (e.g., then, though or however), which may be placed finally (see section 3.3). In written academic prose, however, linking adverbials are almost as frequently placed in various medial positions (40 per cent) as in the default position, the initial position (ca. 50 per cent). This distribution, which has – as has been shown above (cf. section 3.3) – only been common since the Late Modern English period, thus asks for an explanation.

In his 1969 study on adverbial usage, Greenbaum gives a first indication that medial positions of attitudinal disjuncts (stance adverbials) [8] may be triggered by differences in information focus. Greenbaum’s main argument is that disjuncts

are not normally the major information point of the clause to which they are related … however, the attitudinal disjuncts may help to focus the major information points in the clause and this, to some extent at least, accounts for their placement in positions other than the initial position (Greenbaum 1969: 194).

In view of the focus of the present study, it seems worthwhile to repeat some of Greenbaum’s ideas in some detail here:


In the synagogue at Nazareth He [Jesus] significantly read from Isaiah: ….

Analysing the sentence in (12), Greenbaum argues that the disjunct is placed as near as possible to the major information point, viz. ‘the reading from Isaiah’ to focus attention on it (1969: 195). Greenbaum rightly stresses that ‘it would be possible to focus on this part of the clause in a reading even if the disjunct were in initial position’ and contrasts the written form with its spoken counterpart: ‘the reading would achieve this focusing by positioning the nuclear tone on the item with the accompaniment of other prosodic features such as stress or a jump in pitch’ (1969: 195).

In his analysis of (13), Greenbaum (1969) correspondingly argues that appropriately enough places attention on the new element, viz. ‘the first sight by the discoverer of the islets in higher forms’, dividing it from given material:


These ‘follies of Langerhans’ were, appropriately enough, first seen by the discoverer of the islets in higher forms, …

Most interestingly for our present concerns, Greenbaum continues his analysis with the observation ‘Had the disjunct preceded the auxiliary, the subject would have become a major information point’ (1969: 195).

This observation opens up a new perspective on a differentiation of various distinct medial positions: while the disjunct in sentences (12) and (13) highlights the inherent discourse structure of the sentence, i.e., partitions given from new material, its position after the subject is seen to focus attention on the subject. This function of focusing attention on the subject is also illustrated by Greenbaum’s last example where ‘the Subject is focused by a disjunct’, while ‘the additional major information point is graphically indicated by the italicizing of required’ (1969: 195).


Geach, indeed, is required [ital. sic] to say that ‘smokes’ stands for something.

This review of Greenbaum’s examples first of all reveals that the placement of the adverbials, in this case disjuncts, in written English provides an explicit clue to the discourse structure of a sentence, a function which in the spoken mode may rather be effected by different stress patterns or a ‘jump in pitch’. An adverbial in medial position – in particular if set off from the rest of the text by commas – does thus not draw attention to the (unusually placed) adverbial itself, but to other parts of the sentence, i.e., either the immediately preceding element or the following part of the sentence, commonly the rheme. Since this is the default pattern of information structure, the placement of the adverbial here serves to underscore the information structure already present. For these contexts, Taglicht (1984: 22–25) explores the notion of ‘marked rheme’, i.e., a rheme which is given more attention than the usual amount of end focus. Horová (1976: 155) similarly associates this medial position with the transition part of the clause. With the help of the medial adverbial, the rheme is explicitly partitioned from the theme and appears as marked, because it is delayed by an – in all cases – optional adverbial.

4.2.3 Post-initial position

In addition to their function as partitioners, there also seems to be another, somewhat different discourse function of medial adverbials: In addition to ‘right-ward pointing’ adverbials in medial positions which draw attention to the subsequent element as a ‘marked rheme’, there is another ‘medial’ position, which is left-ward pointing, in that it – similar to focus particles such as only or particularly – focuses attention on the preceding element. This is the position after the first sentence constituent (in Quirk et al. 1985 this position is called ‘initial medial position’ – symbol iM; 1985: 490). In my earlier study, I referred to a similar placement – as ‘post-first-position’ or ‘post-initial position’, using a loan translation for German Nacherstposition (see Lenker 2010: 64–72, 235–238). [9]

In German, certain uninflected elements (such as allerdings, wiederum, also, nun, nämlich, beispielsweise) can occur in the “post initial position” between a prefield constituent and the finite verb. […] They are adverbial connectors which, alongside their relational function, in this – and only this – position take on the information-structural task of marking a change of topic. … (Breindl 2008: 27, English abstract)

Let me exemplify the restrictions concerning this position by some examples from Present-Day German (cf. Rehmann will sich nicht länger den Schwarzen Peter zuschieben lassen. Er packte also aus. ‘Rehmann does no longer want to be blamed. So / therefore he blew the whistle’ (Mannheimer Morgen, 30.5.1995; cf. Pasch et al. 2003: 551–552).


Rehmann war der einzige, der zu einer Aussage bereit war. Er also packte aus [post-initial-position possible].

‘Rehmann was the only one who was willing to make a statement. It was him who blew the whistle.’


Rehmann will sich nicht länger den Schwarzen Peter zuschieben lassen. *Er also packte aus. [post-initial-position impossible]

‘Rehmann does no longer want to be blamed. *It was him who blew the whistle’

The comparison of examples (15) and (16) reveals that a post-first-position of the adverbial connector predominantly encodes discourse relations: it is only possible if the element in the first position is in contrastive focus (15). In co-texts such as (16), where the pronoun er ‘he’ is purely anaphoric but not in contrastive focus, the post-first-position of the adverbial connector is not allowed. This contrastive function of adverbial connectors in post-first-position also explains why most of the adverbial connectors which may take this position express the semantic relations of contrast (e.g., Present-Day German aber, allerdings, dagegen, hingegen) or result (e.g., Present-Day German also, schließlich).

The definition of “post-initial position” cited above shows that it cannot be applied to Present-Day English without modification. Since Modern English is not a V2 language, we cannot simply transfer the notions of ‘prefield constituent’ and the ‘finite verb’ position onto English in its younger periods (its applicability on Old English will be tested in a further study). Yet, as the examples from Greenbaum and grammis show, a position after a first constituent seems to fulfil a particular function of focusing attention on the preceding element.

4.3 Discourse functions of adverbial connectors in medial positions

4.3.1 Survey

Combining the findings for the history of adverbial connectors and some first observations about different discourse functions of sentence adverbials (English disjuncts, German adverbial connectors in post-initial position), I would like to suggest here that, because of their form and morphological make-up, adverbial connectors have been a particularly apt means of signalling different discourse functions in the second connect since Late Modern English, when most of them became mono-functional. Those which are not mono-functional in medial position, such as yet (time adverbial or adverbial connector) do not usually occur in medial position when they function as an adverbial connector (cf. also, more, altogether, better, so, etc.; for the full list see Greenbaum 1969: 36–37). More importantly, however, the meaning of adverbial connectors – in contrast to that of adjuncts and disjuncts – in all cases is procedural, not conceptual: They do not, irrespective of the position they take, change the propositional content of any of the segments they relate.

As has been shown above, it is not the adverbial itself which is focused through different medial positions but other constituents in the sentence. In fact, it is one of the defining characteristics of adverbial connectors that their usage underlies a number of restrictions which reflect their functions of focusing attention on other sentence constituents (see Quirk et al. 1985: 631–634 and the full account in Greenbaum 1969: 41–44): They cannot be the focus of a cleft sentence construction (She may be unable to attend the meeting. *It is however that you should send her the agenda.). Similarly, they can neither be the basis of contrast in alternative interrogation or negation (*Should you send her the agenda however or therefore?) nor can they be focused by a focalizer such as only (*You should only <however> send her the agenda). It is the adverbial connectors themselves that fulfil functions similar to these cleft-constructions or focalizers.

These new functions, however, are only a comparatively recent phenomenon. The present-day situation shows a development which only started in the Early Modern English period. [10] This can be nicely illustrated by a corpus-based comparison of the attestations of adverbial connectors in various medial positions on the basis of the complete diachronic corpus compiled for Lenker 2010 (comprising altogether ca. 1.2 million words from Old English to 1920, with an average of about 110,000 words per sub-period; see Lenker 2010: 10–16 and Appendix C). Figure 4 gives the absolute numbers of occurrences of adverbial connectors in various medial positions. This rather fine-grained differentiation is based on Greenbaum’s distinction of medial positions of conjuncts in declarative sentences with a finite verb form (1969: 78). The positions relevant here are

M1between adjunct and subject
M2between subject and verb (including Greenbaum’s (1969) position M3: between subject and auxiliary)
M5between auxiliary and verb (including Greenbaum’s (1969) M6: between auxiliary be and -ed form of verb)
M7between verb be and complement
M8between transitive verb and complement

Figure 4. Mean frequencies (/10,000 words) of adverbial connectors in the various medial positions in the corpus for Lenker 2010: Appendix B).

Figure 4 can only be taken as a first survey, but it corroborates the ideas gained from the analyses of examples (7) and (8). The Late Modern English period saw a drastic rise of the number of medial connectors in academic prose (EModE: 3/100,000 words; LModE 105/100,000 words). While these numbers seem to be highly significant, I refrain from placing final statistical significance onto them: For a comprehensive study of the respective discourse functions, it would be essential to consider all possible realizations of adverbial connectors (i.e., also phrasal and clausal realizations) and to analyse more texts per period. A closer analysis of the corpus texts shows that there is large degree of variation among different authors: The numbers for LModE1, for example, are mainly due to Adam Smith’s predilection for adverbials in M2 position (see example (8)). Yet, the overall picture is clear enough: All of the authors of the Late Modern English period make use of different adverbial connectors in the various medial positions, though with differences in frequency.

What also emerges from the overview in Figure 4, however, is that the position M2, i.e., the post-initial position between the subject and the verb, which was singled out above as having the specific discourse function of emphasizing its preceding constituent, is by far the most frequent one in all periods (57 per cent of all medial connectors in EModE and 48 per cent in LModE). The special character of this position is also attested by the fact that it is similarly not only attested in the Old English period (see Lenker 2010: 70–71), but also found in the Middle English period, which usually avoids medial position of adverbial connectors (Middle English attestations are rare and virtually restricted to inferential thane/then, which often seems to be triggered by Old French dunc; for this see AND, s.v. dunc).

4.3.2 Examples

Since the statistics need further study to be significant and diagnostically conclusive, I will conclude this chapter by giving selected examples for the respective patterns of medial placement of adverbial connectors from a wide range of authors of the Late Modern English period in order to give a first impression of the widespread use of these patterns and their functions in the works of different authors.

M1 and M2

In the discussion above it was suggested that a position after the first constituent – obligatory, i.e., the subject, or optional, i.e., an adverbial – should be singled out as a position which places attention on the preceding element (for the problem of transferring the term Nacherstposition ‘post-initial-position’ coined from V2 Present-Day German to SVO Present-Day English, see section 4.2.3).

M1: between adjunct and subject

In the case of position ‘M1: between adjunct and subject’, the adverbial connector emphasizes the status of the adjunct, which is already marked because of its placement in front position instead of the default position for adjuncts, the end position (see section 4.1 and section 4.2). As has been shown above (section 4.2.1), initially positioned adjuncts fulfil a ‘scene-setting effect’. In (17), the scene-setting temporal adjunct even as late as 1845 is emphasized by the concessive adverbial connector however, which directly follows the adjunct (for a similar example from Smith, see (8f)).


Agardh, … , had observed that, …, the contents of the cells of certain water-weeds were set free, and moved about with considerable velocity, and with all the appearances of spontaneity, as locomotive bodies, which, from their similarity to animals of simple organisation, were called “zoospores.” Even as late as 1845, however, a botanist of Schleiden’s eminence dealt very sceptically with these statements. (CLMET; Huxley, Thomas Henry (1825–1895), Discourses, Discourse VI (1876)).

Similarly in (18), the adjunct in the Muslim world ‘sets the scene’ for the subsequent main clause, which contrasts Muslim attitude with that of Buddhists and Hindus. This contrast is marked by the concessive however in the position between the adjunct and the subject.


How soon organizations arose for the care of the sick, and, in war-time, of the wounded, it would be difficult to say; for Buddhists and Hindus were of course earlier in the field than Muslims, inheriting as they did an older moral culture. In the Muslim world, however, the twelfth century saw the rise of the Kadirite Order, with its philanthropic procedure. (CLEMET; Cheyne, Thomas Kelly (1841-1915), The Reconciliation of Races and Religions, Part 1 (1914)).

M2: between subject and verb

The many examples for the M2 pattern in Smith’s Wealth of the Nations attest this author’s predilection for adverbials (not only adverbial connectors) in this position (see however, therefore, besides and too in examples (8a), (8b), (8e), (8j), (8k)). In all of these cases, the subject, which usually marks a change in topic, is highlighted. This analysis is supported by the finding that ‘impure connectives’, i.e., adverbial connectors which combine their connecting function with a function of contrast/concession or reason/result are frequently found in this position (for the distinction between ‘pure’ and ‘impure connectives’ see Huddleston & Pullum 2002: 775): the most frequent adverbial connectors in this position are resultive therefore (52 instances in EModE and LModE), concessive/contrastive however (42 instances) and inferential then (16 instances). There are very few instances of ‘pure connectives’, which have no other function than that of connecting their clause to the surrounding text, in this position (only too is attested 12 times, while there is not a single instance of post-initial moreover in my corpus). Furthermore, the subjects highlighted in this way by adverbial connectors are almost exclusively lexical noun phrases, often with a demonstrative as a determiner (cf. examples (19) and (20)). In my corpus, there was no instance of an adverbial connector after a subject realized by a purely anaphoric personal pronoun such as he or she (for the similar restrictions to the use of adverbial connectors in Nacherstposition in German, see section 4.3.2).


This point, however, is not essential to my book (CLMET; Chesterton, Gilbert Keith, What’s Wrong with the World, II: The Fallacy of the Umbrella Stand, 1912, III)


Sixty thousand years of HISTORICAL time, when we survey the changes which have come to pass in six thousand, opens to the imagination a range vast enough to seem almost endless. This psychological question, however, need not be decided. (CLMET; Bury, J. B., The Idea of Progress, Preface, Introduction, 1929).

M5, M7 and M8

As has been argued above, all of the other positions of adverbial connectors function as partitioners of given and new information and may even reflect cases of ‘marked rheme’ (see section 4.2.2). This will be exemplified by selected examples for each of the positions. In the patterns M5 and M7, the subject carries the given information, most often also the topical information, while the lexical verb (M5) or the subject complement present the new information (or focus [11]; this part of the sentence is underlined in the following examples). The auxiliary (M5) or the copula (M7) are thus marked to not be part of the new or focus information.

M5: between auxiliary and verb


It may be said that the principles herein advocated, may nevertheless, like the former millions which have misled mankind, originate in error; in the wild and perverted fancy of a well-meaning enthusiasm. They have, however, not only been submitted to several of the most intelligent and acute minds of the present day … (CLMET; Owen, Robert, A New View of Society - Or, Essays on the Principle of the Formation of the Human Character, and the Application of the Principle to Practice, Essay 1, 1813)

M7: between verb be and complement


With a variety of models thus before him, he will avoid that narrowness and poverty of conception which attends a bigoted admiration of a single master, and will cease to follow any favourite where he ceases to excel. This period is, however, still a time of subjection and discipline. (CLMET; Reynolds, Joshua, Seven Discourses on Art, December 11, 1769)


The question of the harlots is, however, quite as insoluble by the ordinary methods. For these unfortunates no one who looks below the surface can fail to have the deepest sympathy. (CLMET; Booth, William, In Darkest England and the Way Out, 1890, Chapter 5)

M8: between transitive verb and complement


I desire, therefore, that you will let me know what is the particular business, department, or province of these several magistrates. (CLMET; Chesterfield, Philip Dormer Stanhope, Letters to his Son(s), 1746–69, Letter II).

5. Conclusions

In the present chapter, I have suggested that there are basically two different ‘medial’ positions of adverbial connectors, namely a so-called ‘post-initial position’ and a genuine ‘medial position’. Both of them serve functions of marking or highlighting information structure in written texts, in which marking of focus by accent or other intonational features is not possible (apart from rare cases of underlining or italization). While adverbial connectors in post-initial position focus attention on their preceding elements, adverbial connectors in other medial positions serve to highlight the partition of given and new information in a sentence and often indicate a ‘marked theme’.

Adverbial connectors are a very apt means for both of these purposes, because their main function is that of marking textual cohesion. In contrast to circumstance adverbials, which may carry important content meaning, they are employed to emphasize a cohesive connection which is already present and are thus the most fitting kind of adverbials to take on further tasks of marking information structure.

This became vital at a time when there was almost no flexibility in word order patterns left. Before the fixation of the word order to SVO, fronting, etc. had been the apt means to highlight information structure also in written texts. Furthermore, only the differentiation of form and function – i.e., the disambiguation of ambiguous adverbs/conjunctions into subordinating conjunctions on the one hand and adverbial connectors on the other – allowed a more flexible placement of adverbials. The stricter rule of punctuation from Late Modern English onwards also made this positional variable more likely to be used. Similar to new cleft-sentences or new passive constructions and a wider flexibility in the kinds of subjects (see Los 2009), the medially placed adverbial connectors became one of the means to indicate thematic structure and textual organization in English, thus helping to compensate the loss of V2 and the fixation of word order to SVO from the Late Modern English period onwards.


[1] I would like to thank Judith Huber, Jonas Bodensohn, Christine Elsweiler, Jutta Öhlschläger, the participants of the Helsinki workshop on connectives and the editor of this volume, Anneli Meurman-Solin, for their most valuable comments on earlier versions of this study.

[2] The term “focus” used as a contrast to “topic” is avoided in this chapter. “Focus” is ambiguous since it may not only refer to the distinction of “topic” and “focus” material, but may also be used as a synonym of ‘emphasis’, ‘attention’. See also endnote 11.

[3] additive and has been used from Old English to Present-Day English. With regard to the contrastive relation, but replaces OE ac from Middle English onwards. For the reformulatory relation the reduced form or (from OE oþþe) has been used since ME3 (for details see Lenker 2010: 76–77). For important differences in the frequencies of these conjunctions in different genres see Meurman-Solin forthcoming.

[4] The sign ° refers to an ‘ambiguous adverb/conjunction’. The only semantic relation specified in Table 2 is that of transition with those adverbial connectors whose connector meaning is not obvious at first glance; all of the other semantic relations are evident from their translations.

[5] The classification used here goes back to one of the most influential monographs on English adverbs by Greenbaum (1969), one of the authors of the Comprehensive Grammar. The 1972 edition of the CGEL corresponds fully to the one described here. The 1985 edition distinguishes a fourth category, namely ‘subjuncts’ (examples: We haven’t yet finished. Would you kindly wait for me? Quirk et al. 1985: 8.88–8.120, pp. 566–612). This fourth category is not adopted here because the whole category of ‘subjunct’ has been repeatedly criticized (see Valera 1998: 267–270). – The Cambridge Grammar of the English Language, however, now suggests another system which mainly affects the boundary between adverb and preposition (Huddleston & Pullum 2002: 264).

[6] Even if the readers’ or hearers’ processing of the discourse is facilitated by the initial linking adverbials, which explicitly mark the logical relationship between the connects, the data also reveal that certain linking adverbials, such as PDE causal then or concessive however and though, may also be placed at the end of a sentence in the spoken medium. In Lenker (2010) this final placement, which is also a comparatively recent phenomenon in English, is also viewed as a consequence of the fixation of word order to SVO (cf. Lenker 2010: 200–213 on sentence-final though).

[7] Most of the studies and grammars follow a common practice and distinguish three main adverbial positions, defined in relation to the verb and obligatory sentence constituents (see Quirk et al. 1985, Biber et al. 1999, Hasselgård 2010). There are differences, however, in the terminology and, in particular, in the numbers of subdivisions which are marked as relevant. Biber et al. (1999: 771) and Huddleston & Pullum (2002: 779) note that there are different medial positions, but do not subdivide them. Quirk et al. (1985: 490, 493–496) differentiate three medial and two end positions; for a similar distinction, see Hasselgård (2010: 41–45).

[8] Greenbaum (1969) does not discuss these patterns for conjuncts, i.e., adverbial connectors.

[9] This position called Nacherstposition is singled out as a specific one by the linguists at the Institut für Deutsche Sprache working on German connectors (see Pasch et al. 2003 and grammis, http://www1.ids-mannheim.de/gra/abgeschlosseneprojekte/konnektoren.html).

[10] The situation in Old English allowed more positions of adverbial connectors than Middle English, but there are – because of the flexibility of word order patterns – other restrictions, which cannot be discussed in any detail in this study.

[11] On the avoidance of the term “focus” in the present study see endnote 2. A closer examination of the examples will, of course, have to distinguish between the “given” and “topic” and “new” and “focus”, which are often congruent, but not fully matching terms. Furthermore, the “theme” vs. “rheme” distinction as taken in the Prague School (Firbas) is also different from the distinction of “given” and “new” (cf. systemic approaches such as Halliday’s), which is more closely tied to the grammar of the clause, i.e. the sequence of clause elements. Thus while “theme” may refer to the point of departure or a kind of topicalization, specifying what the clause is going to be about (and the rheme is what remains), “given” refers to the idea of the speaker which piece of information is recoverable by the hearer from the preceding discourse or the context.

Appendix A: Semantic relations (see Lenker 2010: 39-41)

enumeration and addition (see Lenker 2010: 214–226 and Appendix B.1)

enumeration: firstly, secondly, thirdly (etc.); to begin with


  • equative: correspondingly, equally, likewise, similarly
  • reinforcing: again, also, further, furthermore, moreover, in particular, then, too, (above all)

summation (see Lenker 2010: Appendix B.2)

in sum, to conclude, all in all, in conclusion, overall, to summarize

result/inference/cause (see Lenker 2010: 131–167; Appendix B.3)

therefore, consequently, thus, so; hence, in consequence

contrast/concession (see Lenker 2010: 168–213; Appendix B.4)

on the other hand, in contrast, alternatively; though, anyway, however

  • focus on contrast: conversely, instead, on the contrary, in contrast
  • focus on concession: anyhow, besides, nevertheless, still, in any case

transition (see Lenker 2010: 227–232 and Appendix B.5)

now, meanwhile; incidentally, by the way

Appendix B: Sub-periods (see Lenker 2010: 12)




of Texts

































































Primary Sources

ByrM = Baker, Peter S. and Michael Lapidge (eds.). 1995. Byrthferth’s Enchiridion. (= Early English Text Society. Supplementary Series, 15.) Oxford: Oxford University Press.


CLMET= The Corpus of Late Modern English Texts. http://perswww.kuleuven.be/~u0044428/clmet.htm.


AND = Rothwell, William, Louise W. Stone and Thomas B. W. Reid, eds. 1992. Anglo-Norman Dictionary. (= Modern Humanities Research Association. Publications, 8.) London: Modern Humanities Research Association. Online version: http://www.anglo-norman.net/.

MED = Kurath, Hans, Sherman M. Kuhn, Robert E. Lewis et al., eds. 1954–2001. Middle English Dictionary. Ann Arbor, Michigan: University of Michigan Press. Online version: Middle English Compendium, http://ets.umdl.umich.edu/m/mec/.


Bækken, Bjørg. 1998. Word Order Patterns in Early Modern English with Special Reference to the Position of the Subject and the Finite Verb. (= Studia Anglistica Norvegia, 9.) Oslo: Novus.

Biber, Douglas, Stig Johansson, Geoffrey Leech, Susan Conrad & Edward Finegan. 1999. Longman Grammar of Spoken and Written English. London: Longman. [= LGSWE]

Campbell, George. [1776] 1963. The Philosophy of Rhetoric, ed. by Lloyd F. Bitzer. Carbondale: Southern Illinois University Press. [First published London: Printed for W. Strahan and T. Cadell and W. Creech at Edinburgh 1776.]

Breindl, Eva. 2008. “Die Brigitte nun kann der Hans nicht ausstehen. Gebundene Topiks im Deutschen”. Deutsche Sprache 36: 27–49.

Couper-Kuhlen, Elizabeth & Bernd Kortmann, eds. 2000. Cause – Condition – Concession – Contrast. (= Topics in English Linguistics, 33.) Berlin: de Gruyter.

Culpeper, Jonathan & Merja Kytö. 2000. “The conjunction and in Early Modern English: Frequencies and uses in speech-related writing and other texts”. Generative Theory and Corpus Studies. A Dialogue from 10 ICEHL (=Topics in English Linguistics, 31), ed. by Ricardo Bermúdez-Otero, David Denison, Richard M. Hogg & Chris B. McCully, 299–326. Berlin: Mouton de Gruyter.

Diessel, Holger. 2005. “Competing motivations for the ordering of main and adverbial clause”. Linguistics 43 (3): 449–470.

Ford, Cecilia E. 1993. Grammar in Interaction: Adverbial Clauses in American English Conversations. (=Studies in Interactional Sociolinguistics, 9.) Cambridge: Cambridge University Press.

Givón, Talmy. 2001. Syntax: An Introduction, vol. II. Amsterdam: Benjamins.

grammis = grammis. das grammatische informationssystem des ids. Mannheim: Institut für deutsche Sprache. http://hypermedia.ids-mannheim.de/.

Greenbaum, Sidney. 1969. Studies in Adverbial Usage. London: Longman.

Haiman, John & Sandra A. Thompson, eds. 1988. Clause Combining in Grammar and Discourse. (=Typological Studies in Language, 18.) Amsterdam: Benjamins.

Hasselgård, Hilde. 2010. Adjunct Adverbials in English. Cambridge: Cambridge University Press.

Horová, Eva. 1976. “On position and function of English local and temporal adverbials”. Brno Studies in English 12: 93–123.

Huddleston, Rodney & Geoffrey K. Pullum. 2002. The Cambridge Grammar of the English Language. Cambridge: Cambridge University Press.

Jacobson, Sven. 1964. Adverbial Positions in English. Stockholm: Studentbok.

Kemenade, Ans van & Bettelou Los. 2006. “Discourse adverbs and clausal syntax in Old and Middle English”. The Handbook of the History of English, ed. by Ans van Kemenade & Bettelou Los, 224–248. Oxford: Blackwell.

Kemenade, Ans van, Tanja Milicev & R. Harald Baayen. 2008. “The balance between syntax and discourse in Old English”. English Historical Linguistics 2006. Selected Papers from the Fourteenth International Conference on English Historical Linguistics (ICEHL 14), Bergamo, 2125 August 2006, Vol. I: Syntax and Morphology (= Current Issues in Linguistic Theory, 295.), ed. by Maurizio Gotti, Marina Dossena & Richard Dury, 3–21. Amsterdam & Philadelphia: Benjamins.

Kerkhof, Jelle. 1982. Studies in the Language of Geoffrey Chaucer. (=Leidse germanistische en anglistische reeks, 5.) Leiden: Brill.

Kohnen, Thomas. 2007. “Connective profiles in the history of English texts: Aspects of orality and literacy”. In Lenker & Meurman-Solin (eds.), 289–306.

Kortmann, Bernd. 1997. Adverbial Subordination. A Typology and History of Adverbial Subordinators Based on European Languages. (=Empirical Approaches to Language Typology, 18.) Berlin/New York: Mouton de Gruyter.

Lenker, Ursula & Anneli Meurman-Solin, eds. 2007. Connectives in the History of English. (=Current Issues in Linguistic Theory, 283.) Amsterdam: Benjamins.

Lenker, Ursula. 2010. Argument and Rhetoric – Adverbial Connectors in the History of English. (=Topics in English Linguistics, 64.) Berlin: Mouton de Gruyter.

Los, Bettelou. 2009. “The consequences of the loss of verb-second in English: information structure and syntax in interaction”. English Language and Linguistics 13: 97–125.

Meurman-Solin, Anneli & Päivi Pahta. 2006. “Circumstantial adverbials in discourse: A synchronic and a diachronic perspective”. The Changing Face of Corpus Linguistics, ed. by Antoinette Renouf & Andrew Kehoe, 117–141. Amsterdam and New York: Rodopi.

Meurman-Solin, Anneli. Forthcoming. “The connectives and, for, but, and only as clause and discourse type indicators in 16th- and 17th-century epistolary prose”. Information Structure and Syntactic Change in the History of English, ed. by Anneli Meurman-Solin, María José López-Couso & Bettelou Los (=Oxford Studies in the History of English). New York: Oxford University Press.

Mitchell, Bruce. 1985. Old English Syntax, 2 volumes. Oxford: Clarendon.

Mueller, Janel L. 1984. The Native Tongue and the Word. Developments in English Prose Style 1380–1580. Chicago: University of Chicago Press.

Pasch, Renate, Ursula Brauße, Eva Breindl & Ulrich Hermann Waßner. 2003. Handbuch der deutschen Konnektoren. Linguistische Grundlagen der Beschreibung und syntaktische Merkmale der deutschen Satzverknüpfer (Konjunktionen, Satzadverbien und Partikeln). Berlin: de Gruyter.

Parkes, Malcolm B. 1992. Pause and Effect. An Introduction to the History of Punctuation in the West. Aldershot: Scolar Press.

Quirk, Randolph, Sidney Greenbaum, Geoffrey Leech & Jan Svartvik. 1985. A Comprehensive Grammar of the English Language. London: Longman.

Quirk, Randolph & Charles Wrenn. 1977. An Old English Grammar (Methuen’s Old English Library). London: Methuen.

Svenonius, Peter. 2002. “Subject positions and the placement of adverbials”. Subjects, Expletives, and the EPP, ed. by Peter Svenonius, 201–242. Oxford: Oxford University Press.

Swan, Toril. 1988. Sentence Adverbials in English: A Synchronic and Diachronic Investigation. (=Tromsø-studier i språkvitenskap, 10.) Oslo: Novus.

Taglicht, Josef. 1984. Message and Emphasis. London: Longman.

Thompson, Sandra A. 1985. “Grammar and written discourse: Initial vs. final purpose clauses in English”. Text 5(1): 55–84.

Ungerer, Friedrich. 1988. Syntax der englischen Adverbialien. Tübingen: Niemeyer.

Ungerer, Friedrich, Gerhard E. H. Meier, Klaus Schäfer & Shirley B. Lechler. 1984. A Grammar of Present-Day English. Stuttgart: Klett.

Valera, Salvador. 1998. “On subject-orientation in English ly-adverbs”. English Language and Linguistics 2(2): 263–282.