Methodological suggestions for investigating Shakespearean discourse markers in old texts of Shakespeare’s plays [1]

Beatrix Busse, Heidelberg University
Ulrich Busse, Martin Luther University Halle-Wittenberg

Abstract

1. Introduction and research questions

The underlying idea of the present paper is to illustrate how the investigation of Shakespearean discourse markers in old texts of Shakespeare’s plays can be fruitful for the retrieval of historical forms and for determining how small adjustments in the use of discourse markers add to subtle changes of meaning. We will embrace a functional approach to language, which sees each choice as meaning-making and stylistic and must hence address the context of text production (see B. Busse 2010b). Here, we explore the interplay between editorial considerations and pragma-linguistic questions (but we will not address questions about the “best” text). In particular, we will focus on the relationship between old texts and early printings and the choices of discourse markers displayed in them.

Blake (2002: 293) illustrates the need for acknowledging how editorial considerations and what is of interest to historical pragmatics may coincide. He shows how, for example, in the First Folio (F1) from 1623, in other quarto editions, and in modern editions of respective Shakespeare plays functions of words like well or say have been interpreted differently by editors. He discusses example (1) below. In The Merry Wives of Windsor (Wiv.). Page uses the formula I am glad to see you frequently. The first occurrence of this phrase in F1 reads as follows:

“Modern editors keep well in this position as an adverb” (Blake 2002: 293). Thus, in the Riverside Shakespeare it is treated accordingly:

But Blake argues that it could also be seen as a discourse marker introducing the following clause. This interpretation would give rise to the theoretical punctuation of this line in (1.b):

Blake offers convincing reasons for the interpretation shown in (1.b). While in Present-day English (PDE) well collocates with other discourse markers, such as you know, this was uncommon in Early Modern English (EModE). In Shakespeare’s English, well “is most frequently associated with forms of address in the quarto: Well sirs, (Wiv. 1. 3. 27, Q Well my Laddes), many of which do not occur in F” (Blake 2002: 293). Therefore, the interpretation of an item like well clearly depends upon the choice of the copy-text, i.e. Quarto text(s) or Folio text. But even this leaves room for editorial intervention.

When going back to the historical moments of text production a considerable number of manuscripts and old texts of Shakespeare’s plays can be investigated. However, despite the fact that these old texts reveal to us important diachronic clues as to the pragmatic and meaning-making linguistic choices made at the time when Shakespeare composed his plays and beyond, it takes some expert knowledge of editorial theory and textual criticism to understand, for example, an editorial apparatus of a critical or variorum edition of a Shakespeare play (see Taavitsainen and Fitzmaurice 2007: 21). In addition, whether linguistic items such as well function as an adverb or rather as a discourse marker in a given historical context of old texts of Shakespeare’s plays “is sometimes difficult to tell” (Blake 2002: 293), especially when their occurrences are compared with each other in the existent historical printings (within a particular time-frame).

Editors can be seen as mediators of the past (see B. Busse 2010b), because they make particular choices on the basis of the copy-text. These, in turn, must influence historical-pragmatic interpretations and demand of the historical pragmatician to be clear about the text – old or modern – under investigation. This is stressed by Taavitsainen and Fitzmaurice (2007) in their programmatic introduction “Historical pragmatics: What it is and how to do it” to the volume Methods in Historical Pragmatics:

The process of editing involves decisions between alternatives from blurred and indistinct letter forms that might invite different interpretations to expanding abbreviations, marking corrupt forms and possible emendations. (Taavitsainen and Fitzmaurice 2007: 21)

In historical pragmatics, there is no lack of awareness of the complexity of analysing historical texts and manuscripts:

The linguistic study of historical texts is complicated. It is natural to encounter obscurity, vagueness, and ambiguity of language use in the material products of distant cultures with no direct access to the speakers and original contexts of production. (Taavitsainen and Fitzmaurice 2007: 11)

The extent to which the choice of old and modern texts and textual editions of Shakespeare’s plays – both quantitatively and qualitatively – affects our pragmatic analyses and interpretations will most likely remain difficult to determine on a large scale. Therefore, the present paper addresses the following question: To what extent does a historical collation of the use of selected discourse markers in old texts of Shakespeare’s plays provide us with additional information about historical language usage?

We agree with Taavitsainen and Fitzmaurice (2007: 21) that “return[ing] to original manuscript sources” alone and no longer using modern editions cannot be a panacea for the complex challenge of accounting for an historically informed and valid interpretation of both old texts of Shakespeare’s plays or of pragmatic features in general. We are far away from arguing this, because determining the differences between existing old texts and modern textual editions has been admirably shown by studies of editorial theory and textual criticism. Furthermore, an analytical focus on old texts of Shakespeare’s plays is not a different study of Shakespeare’s plays altogether that will reveal a totally different outcome. Following a long tradition in textual editing, each (modern) edition of Shakespeare uses as its copy-text – that is, as the text that is the basis of the edition – a specific old text, i.e., a good quarto, or in the case of the F1-only plays the First Folio (F1) from 1623. We are aware of the fact that modern editions of Shakespeare’s plays are not only based on a specific copy-text, but that virtually all modern editors consult other available versions, too. However, what we would like to argue is that making use of the (often rare) situation to have more than one historical text of a Shakespeare play opens up an additional methodological component of performing an historically informed investigation of discourse markers (see B. Busse 2010b). We will therefore show that the investigation of discourse markers in old texts of Shakespeare’s plays reveals highly meaning-making information about the use of pragmatic phenomena in Shakespeare’s plays and in EModE in general.

Due to our focus on one particular pragmatic function – discourse markers – and their respective forms, it is possible to make interpretations on the basis of some quantifiable tendencies. As a methodological procedure, the comparison of old texts for their usage of discourse markers has also not been extensively studied from a historical pragmatic perspective – one reason being that the textual situation is not always as well documented as in the case of Shakespeare. Also, despite the fact that scholars of editorial studies and textual criticism, who have produced outstanding (variorum) editions of Shakespeare’s plays and other texts, would have editorial expert knowledge for us to learn from – for example, how to use and understand a text-critical apparatus – cooperation between editors and linguists has been scarce. [2] Our cooperation with Marga Munkelt (Münster, Germany), who is a textual editor and who worked with Marvin Spevack on the concordances, has helped us to understand the complicated textual situation with regard to old texts of Shakespeare’s plays.

This then raises the question of what textual sources are available. Investigations of the editorial choices and the way these can be revealing for the analysis of discourse markers need to draw on either well-documented modern editions of single plays which contain a complex diachronic text-editorial apparatus, on variorum editions of single plays or on (electronic) Shakespeare corpora which diachronically display the variants of the different textual traditions. [3] As the Shakespeare Database, compiled under the supervision of H. J. Neuhaus at the University of Münster (Germany), is still not publicly available, although it would have at least in parts provided us with exactly this kind of information, we chose Volume IX (1980) of Spevack’s (1968–1980) computer-generated Shakespeare Concordance. [4]

Volume IX of the concordance is of special interest for our purpose, since it is “a collation of substantive variants [in old printings], recording the accepted readings in eight major [modern] editions from the second Cambridge Shakespeare (1891–93) through The Riverside (1974)” (Evans 1974: 69). In other words, Volume IX lists all those words and passages which are different in the respective old texts and which do not occur in the copy-text of the modern editions and therefore cause a deviation from the respective copy-text to such an extent that they change the meaning of the respective line. This includes, for example, lexical, syntactic and also orthographic deviations. We therefore consider Volume IX to be a corpus, a corpus of variants. This corpus contains two subcorpora:

As such, volume IX of Spevack’s (1968–1980) concordance can also be used for studies going beyond Shakespeare – because it lists linguistic forms of particular periods. However, for our pilot study we only studied the variants for particular discourse markers of the old texts systematically.

By drawing on Volume IX of Spevack’s (1968–1980) Shakespeare Concordance we investigate discourse markers by means of the following methodological triangulation: We start by examining whether discourse markers are affected – and if so how – when textual passages are altered in a given play. To this end, we studied Romeo and Juliet (Rom.) as a test case because it has got an interesting textual situation with three existing old printings. [5] In a second approach, we investigate a specific form of a discourse marker, well, for those plays, which according to Volume IX of the Spakespeare Concordance show considerable variation in old (and new) editions. Finally, we deal with discourse markers which have been identified by B. Busse (2010a, 2012) as stance adverbials in the entire corpus of Shakespeare’s plays.

Our aim is to show that with the help of Volume IX (1980) of the Shakespeare Concordance (Spevack 1968–1980) it is possible to return to manuscripts even in historial pragmatic research and to do this on a larger scale. Furthermore, we would like to illustrate that the scrupulous study of variants in old texts and editions of Shakespeare’s plays is not just old-fashioned philological nit-picking and crux-busting but reveals hitherto unexplored aspects of historical usage of discourse markers, which has qualitative and quantitative consequences for how pragmatic phenomena may be interpreted.

For Shakespeare, there is consensus that there are major old texts (Quarto – the so-called good Quartos – and F1), but there are different traditions as to which texts are seen as copy-texts for the individual plays. As a very general rule of thumb – and admittedly slightly simplified – it can be said that if there are quarto editions of a play, usually the copy-text will be one of the good quarto texts. The quarto texts can also be regarded as “theatre versions”. They are close to the theatrical tradition, because they were produced after a performance took place. By contrast, F1 from 1623 can be seen as a “first literary piece” in a particular tradition. There are 18 F1-only plays (which do not have quarto editions).

We study those words/passages in the text, which are different in the existing old printings and which do not occur in the copy-text of the modern editions. We would like to answer the following questions:

• What discourse markers are used in the different versions of the texts available?
• What are the results if the different versions are compared to each other?
• What additional information do we get about the status of discourse markers in Shakespeare’s plays by doing this?

These questions are valid with regard to two aspects:

Apart from the different traditions and practices, important social changes took place between the printing of the quartos and F1 as well. They also affected the language used in the plays as they are transferred to us through the different old texts. For instance, the 1606 Profanity Act “An Act to Restrain Abuses of Players” is very important for our investigation and the linguistic status of F1. The act prohibited spoken profanity, irreligion in any dramatic production, policing of speech with religious reference and resulted in textual censorship of F1 (Gazzard 2009: 495f.). This act therefore clearly affected the chosen forms of discourse markers (as shown in Section 3.3) as displayed in F1, when, for example, by god was replaced by heaven in F1. While this linguistic censorship has been studied (Gazzard 2009) from an editorial philological point of view, it is important for us to analyse from a pragmatic point of view which forms have been chosen to replace those that contained religious reference in editions preceding F1.

2. A working definition of discourse markers

Discourse markers have received much attention in the wider field of historical pragmatics over the past decades. [6] However, with exclusive regard to Shakespeare’s English this topic is much less charted than, for example, address terms or politeness (see Busse and Busse 2010: 266–268). This alone warrants further investigation, and the methodological challenges outlined in Section 1 make this topic particularly interesting for “developing corpus methodology for historical pragmatics” – to quote the title from the pre-conference event to the Helsinki Corpus Festival held in Helsinki 28 September – 2 October 2011.

Discourse markers do not constitute a formal grammatical class […] since they encompass one-word items such as actually, anyway, like, now, so or well – which may be homophonous with adverbs, prepositions or conjunctions – but also phrasal elements such as in fact, after all or and stuff like that, and abridged clauses such as you know, I mean or you see. (Brinton 2010: 285)

For the EModE period, Brinton (2010: 290–292) documents the following “new” discourse markers and categorises them according to their internal complexity ranging from one-word items to comment clauses (which we have arranged in alphabetical order in Table 1 below).

Out of the inventory as shown in Table 1, we made a small selection and approached the problem from two different angles: starting with a given play (outlined in Section 3.1) and then moving on to a given discourse marker (as shown in Section 3.2). In addition to the discourse markers presented in Table 1) we added the group of stance adverbials (see Section 3.3).

3. Corpus study of selected Shakespearean discourse markers

As mentioned in Section 1., we used Volume IX of the Shakespeare Concordance (Spevack 1968–1980) as our corpus and looked for textual variation with regard to selected discourse markers. A study of the concordance lines soon made clear that this approach only works for plays with a difficult transmission history, because only such plays provide enough examples of variants.

3.1 Romeo and Juliet as a test case

We chose Rom. because it fulfils these criteria. The play has an interesting textual history with three existing early printings: Q1, Q2, and F1, with Q2 being the copy-text, also in The Riverside Shakespeare. Evans (1997) outlines the complicated transmission history of the play as follows:

Romeo and Juliet presents a complicated textual problem. The play was first published in a “bad” quarto version (Q1) in 1597. A “good” quarto (Q2), “Newly corrected, augmented, and amended,” appeared two years later in 1599. Three other quarto editions were printed, each set up from the immediately preceding edition: Q3 (1609), Q4 (undated, but probably c. 1622), Q5 (1637). The First Folio text (1623) was set directly from a copy of Q3, with almost no attempt at correction apart from the addition of a few obvious stage directions. (Evans 1997: 1139)

Even though Volume IX of the Shakespeare Concordance lists only a few variants of discourse markers in Rom., some revealing tendencies can be observed nonetheless; i.e., whether discourse markers feature in a particular textual passage or not and what effect this produces when textual passages are altered. The discourse markers showing variation are ah (see Examples (2) and (3)), how how / how now (4) and by my troth (5).

This example is difficult to interpret. On the one hand, it can illustrate a change from the indefinite article into a discourse marker in the modern editions, but, on the other hand, “a” could also be a spelling variant of “ah” as in the OED definition for “a” interjection: “†1. Expressing invocation, surprise, or admiration; (also) used to gain attention. Obs.” (with a last citation from “around 1500”). So this could be a spelling change or a wordclass change. Meaning would suggest a spelling change.

The discourse markers “ah” and “o[h]” differ in their illocutionary force: “Againe, for expressing our passions, our interiections are very apt and forcible: as findeinge ourselues somewhat agreeued, wee cry Ah; yf more deeply, Oh […]” (Carew 1595/96 quoted in Kühlwein 1971: 39)

Example (4) shows that discourse markers occur in all old printings, but there are different forms. So our question would be whether there is a difference in meaning between “how, how!” and “how now!”

Example (5) illustrates: Whoever wrote down the text in Q2 might have misheard “by my” or it is a printing error, because “by my” makes more sense. In any case, by my troth is a fixed phrase as illustrated in Section 3.3.

The analysis of these few examples reveals that the variation between the old texts and between the old and modern texts can affect the meaning of a given passage. Some of the cases above may not be “substantive variants”. They could be spelling variations or just printer’s errors. However, we believe that these cases do not weaken the case that variant analysis can lead to significant findings, but the “noise” in the data shows that a concordance of “substantive variants” can only serve as a corpus, and that it still takes human interpretation to find out whether a variant is “substantive”, uninteresting or just an error.

Since the results for a single play are not generalisable or quantifiable on a larger scale, we supplemented this search by a second approach.

3.2 Well as a test case

The second approach does not take a given play as its starting point, but a discourse marker. Here, we used well as a test case. Well occurs frequently in Shakespeare’s plays and it has been discussed previously (see Jucker 1997). We looked whether Volume IX lists well as a “substantive variant”. All in all, we found 32 variants for well in all of the plays. Depending upon the transmission history of a given play, the number of tokens varies considerably. The two plays Richard III (R3) and Othello (Oth). were quite rich in textual variants concerning well. In the following, R3 will be discussed in some detail. The textual transmission of R3 looks like this:

Richard III presents a difficult textual problem. The play first appeared in quarto (Q1) in 1597, and five more quarto editions, Q2 (1598), Q3 (1602), Q4 (1605), Q5 (1612), Q6 (1622), were printed before the publication of the First Folio (1623) text. Each of the quartos after Q1 was printed from the immediately preceding edition (Q5 from both Q3 and Q4), each new edition compounding the errors of its predecessor and adding new ones of its own. Two later quartos, published in 1629 (Q7) and 1634 (Q8), are textually of no concern here [i.e. in the Riverside Shakespeare]. (Evans 1997: 794) [8]

The results obtained for well used as a discourse marker in R3 are illustrated and discussed in Examples (6) to (12) below.

The text shown in (6) is from F1, which also served as the copy-text for the Riverside Shakespeare (1974) edition of the play. Seven major modern and representative editions as recorded in the concordance have followed this text – the only exception being CAM2. [9]

In contrast to this, all the quarto editions from Q1 to Q6 and also CAM2 present a slightly different text, in which now occurs instead of well, as shown in (6.a) below.

This example shows that the adverb now used in the quarto editions, was replaced by well in F1 and all major modern editions. This is a revealing pattern because the Quarto editions, which are closer to the theatrical production, use the temporal adverb now in collocation with the modal auxiliary must to transfer the urgency of the task. The use of well in F1 is more jocular and light-hearted in tone. In contrast to Murderer 2, Murderer 1 does not repent their murder and makes fun of Murderer 2’s cowardness. Therefore, for him it is easy to “hide the body in some hole”.

Example (7/7.a) shows that well in F1 replaces the comment clause I tell thee used in the earlier quarto editions. Again, it is tempting to argue that the reduction in this line is most likely made for the rhythm to create the iambic pentameter line (with a feminine ending). However, the reduction could be due to the fact that F1 reduces the length of the discursive/pragmatic element, and that at the time well was more frequently used than the comment clause I tell thee. In essence, one discourse marker was replaced by another one.

Example (8/8.a) illustrates that a discourse marker which was present in the earlier quarto editions was deleted in F1. Several explanations are possible for this change: In terms of pragmatics, the illocutionary force of the utterance is changed with the deletion of well because “Call them again […]” without the discourse marker well is a straightforward imperative that is not downplayed by a discourse marker. In other words, “Well, call […]” sounds more insecure than the imperative “Call them  again […]”.

With regard to aesthetic considerations one could argue that the F editors try to adhere to the iambic pentameter because the Q1 line contains one too many syllables. In addition, the eliminating of “well” also has the effect of strengthening the alliteration between “cares” in the previous line and “calls”. This interpretation underlines the argument that F is the more literary text.

In comparison to Example (8/8a) Example (9/9.a) works the other way around, well is absent in all the quarto editions, but was inserted in F1. One reason might be that the F editors were correcting for rhythm, either by adding or substracting syllables to get the unrhymed iambic verse right.

In (10/10.a) – well replaces earlier saie. Again, it seems that a discourse marker originating from the comment clause I say ... was replaced by a “shorter” form. But that would have to be verified, of course, on the basis of more forms. There is also a difference in illocutionary force when the two forms well and saie are compared. “Well, then” creates assonance in the first two words and contains a hint of sarcasm in Elizabeth’s utterance, and no longer the power of the command from the imperative in saie. She pretends she does not understand who will be the future husband of her daughter, even though Richard has informed her that it will be him.

Somewhat contrasting to the examples quoted before is Example (11/11.a). The discourse marker well is replaced by then in F1. Here rhythm is not an issue. In examples (10/10a.), Elizabeth is characterised by means of her use of the discourse marker well to be a little less authoritative, but sarcastic in the F1 version, where well is introduced. Well present in the quarto editions in (11.a) adds a different illocutionary force because the topic that is talked about is different. Richard needs to collect his men. He seems to be weaker and hesitating in the Q version with “Well, go”, which is also interrupted by the use of the comma. In the F version, he is much more powerful and immediately gives his command. Then and men in F create assonance.

Example (12/12.a) shows again that a discourse marker which was not present in any of the earlier quarto editions was introduced in F1 because of rhythm.

In summary, these examples illustrate the following tendencies: F1 clearly played the determining role for most modern editions, in that, excepting CAM2, all the modern editions have followed F1. A comparison of the treatment of well in the quarto editions and F1 shows that well may be omitted, inserted or replaced by other discourse markers or adverbials.

One explanation may be to even out the rhythm. In some cases the changes in F1 strengthen (or weaken) characters. In this respect, the changes serve a double function: they are pragmatic (in that they change the illocutionary force of an utterance), and at the same time they are literary (in that they bring about subtle changes in characterization). Taken together, both argumentations (rhythm and pragmatics/characterization) work hand in hand in that they clearly underline that F1 is the more literary text.

3.3 Stance adverbials as test cases

B. Busse (2010a, 2012) investigates the forms, frequencies and functions of stance adverbials in a selection of Shakespeare’s plays, which have as a textual basis The Riverside Shakespeare (Evans 1997). Stance adverbials indicate a speaker’s attitude or opinion. They can be defined as sentence modifiers with sentential scope, which are not grammatically required but may be pragmatically desired and are speaker- and/or hearer-oriented. Hence, they are sometimes seen as discourse markers. According to Biber et al. (1999), they can be (a) epistemic, (b) attitudinal, and (c) style adverbials. Examples of epistemic stance adverbials are indeed (doubt and certainty) or in regard of (viewpoint). Attitudinal stance adverbials express the speaker’s attitude towards an evaluation, as in haply, for example. Indeed in “Here’s a change indeed! (Oth. 4. 2. 106, our emphasis), or to say precisely in “And indeed such a fellow, to say precisely, were not for the court (All's Well That Ends Well 2. 2. 12, our emphasis) or by my life in “Helen, I love thee, by my life I do” (A Midsummer Night's Dream 3. 2. 51, our emphasis) illustrate the use of stance adverbials in Shakespeare. B. Busse (2010a, 2012) shows that indeed, truly, faith, in faith and by my troth are the most frequently occurring forms of stance adverbials in Shakespeare’s plays. In addition, the construction by (my), which is followed by a number of nouns, such as faith, the lord, my mother, my white beard and so on, is the most frequently occurring construction in Shakespeare’s dramatic work. On the basis of this quantitative distribution, we investigate variants containing the by (my) construction in Volume IX of the Shakespeare Concordance. All in all, there are 46 cases and most of these occur in the histories (39), especially Henry IV, Parts 1 and 2 (1H4 and 2H4). The tragedies contain five examples and there are two examples from the comedies.

A number of patterns can be observed. Generally speaking, F1 usually omits this construction when, in the chronologically earlier quarto editions, the semantics of the construction used refers to a religious semantic field, that is, to God or to one’s faith.  Especially by my faith, by the lord, by the mass and sometimes also by my troth – that is the frequent pattern for these forms – are ommitted in F1. One reason, which is also highlighted by Gazzard (2009), is the already mentioned Profanity Act from 1606, in which blasphemy in the theatre is sanctioned. Hence, the lack of those forms in the F1 edition.

Example (13), in which by the lord occurs in the quarto editions, but not in F1, illustrates this development. The modern Riverside Shakespeare (Evans 1997) edition, whose copy-text is F1, introduces a variant here by drawing on Q1 and re-using the stance marker by the Lord.

The same holds true for example (14):

In Example (15), we can see that the construction by my troth is omitted in F1:

But there is deviation from this pattern, too. While the preceding examples illustrate that stance markers with reference to a religious semantic field are simply ommitted in F1, the following examples show that these can also be replaced by either interjections or other stance adverbials realised through comment clauses. These cases provide us with important historical-pragmatic information because they a) list forms of discourse markers that are used around 1623 and b) allow us to compare the choices made in the Quarto editions with those in F1. Example (16) illustrates that the attitudinal construction by the mass is replaced by the interjection look, look.

In Example (17), the epistemic stance adverbial by the mass, which indicates – albeit in a rather formulaic way – the certainty of what is said, is changed into an initially placed comment clause with it seems, which, despite the fact that the religious reference is deleted, strikingly marks uncertainty instead and therewith also adds to the characterisation of Hamlet being more in doubt than in the Q2 version.

We have found examples in which stance adverbials indicating sincerity can also be replaced by other stance adverbials fulfilling the same functions. In Examples (18) and (19), by my troth is replaced by the shorter and, at the time, even more frequently occurring stance adverbial tru(e)ly, despite the fact that by my troth is a rather formulaic form at Shakespeare’s time and beyond. Conventionally speaking, both indicate sincerity.

Example (20) is equally revealing, because the replaced form represents a construction that, at Shakespeare’s time, does not occur as frequently (B. Busse 2012): by my faith is replaced by to speak truth, which consists of a verb of communication, followed by a noun.

Similar trends can be observed for the stance marker indeed, which is the most frequently occurring form marking attitude in Shakespeare’s plays (B. Busse 2010a, 2012), but which does not have so many variants listed in the concordance when starting from a form-to-function approach. Stance markers which have a religious semantic reference in the quarto editions are replaced by indeed in F1. Example (21) gives evidence of this.

All variant examples of the construction by (my) and indeed and the ways how they are omitted or replaced show the relevance of discourse markers, such as stance adverbials, in general, and those epistemic stance adverbials indicating sincerity, in particular. Socio-political developments such as the Profanity Act from 1606 affect and construe language usage. At the same time, the visibility of this kind of censorship in F1 from 1623 is highly useful for a historical pragmatician, because it tells us which forms have already undergone a process of grammaticalisation, and which forms then have been chosen to serve as epistemic stance markers indicating both sincerity or insincerity.

4. Conclusion and outlook

The interplay between pragmaphilology, diachronic pragmatics and aspects of textual and editorial theory is fruitful. Editors have already drawn attention to the ways in which variants and old printings interact with one another. Historical pragmaticians need to take account of this and systematically investigate the changes made in different editions, because from a diachronic point of view these give highly useful pragmatic clues about language usage at a particular historical moment.

The analysis of our various test cases has shown that seemingly minute details can affect a pragmatic interpretation of discourse markers. We found differences or even tendencies between the quarto edition(s) and F1 and the interesting shifts occur when modification patterns shift, and where that shift indicates a change in meaning. However, the sample was not large enough to spell out different preferences, practices or even styles in the quartos and in F1. Further research must investigate a set of discourse markers systematically on a larger scale. Volume IX of Spevack’s (1968–1980) Shakespeare Concordance will play a crucial role. Our case study has shown that it is still an invaluable research tool.

Notes

[1] We would like to thank the anonymous reviewers for their instructive comments and suggestions.

[2] Among the many modern textual editors, who have done so much to problematize our understanding of what a Shakespeare text really is, Stanley Wells and Gary Taylor, Ann Thompson, Leah Marcus and a host of others need to be mentioned.

[3] It is possible to view and compare 107 copies of 21 plays by Shakespeare printed in quarto before 1642 at the British Library's Treasures in Full website. Though this is not the equivalent of a corpus used for systematic linguistic studies, it does give some sense of changing editorial practices in the early modern period.

[4] On data and methodology for Shakespeare studies see also Busse and Busse (2010: 249).

[5] It should be stressed here that Rom. is certainly not the only Shakespeare play with an interesting textual situation. Hence, Rom. serves as an exemplary text case.

[6] For a recent overview see Brinton (2010).

[7] The relevant lines in this and all the following examples were identified by means of Volume IX of the Shakespeare Concordance (Spevack 1968–1980). The text is quoted according to Evans (1997) The Riverside Shakespeare. All departures of this edition from the copy-text are listed in detail in the section “Note on the text” after each play.

[8] As regards the textual basis of F1, Evans provides the following information: “F1 offers a similar ambiguous authority, since it was printed from copies of Q3 and Q6 […], which had been corrected and augmented against an independent manuscript, either Shakespeare’s ‘foul papers’ or a scribal copy of the ‘foul papers,’ but in any case almost certainly not a manuscript, such as a prompt-book, with direct theatrical connections. Two stretches of the F1 text (III. i. 1–158) and V. iii. 48 to the end of the play),printed from Q3, show no evidence of any correction against such a manuscript” (1997: 794).

[9] The modern editions of CAM3, KIT1, ARD2, ALEX, SIS, PEL2, EVNS have followed this text. For details of the procedure for recording these variants see (Shakespeare Concordance Volume IX, 1980: vii–xiii).

References

1. Corpus

Evans, G. Blakemore. 1974 & 1997. The Riverside Shakespeare. Boston: Houghton Mifflin.

Spevack, Marvin. 1980. Substantive Variants. Vol. IX of A Complete and Systematic Concordance to the Works of Shakespeare (1968–1980). Hildesheim: Olms; 9 vols.

2. Secondary Sources

Biber, Douglas, Stig Johannsson, Geoffrey Leech, Susan Conrad, & Edward Finegan. 1999. Longman Grammar of Spoken and Written English. London: Longman.

Blake, Norman F. 2002. A Grammar of Shakespeare’s Language. Houndmills: Palgrave.

Brinton, Laurel J. 2010. “Discourse Markers”. Historical Pragmatics, ed. by Andreas H. Jucker & Irma Taavitsainen, 285–314. Berlin: Mouton.

Busse, Beatrix. 2010a. “Adverbial Expressions of Stance in Early Modern ‘Spoken’ Language”. Anglistentag 2009 Klagenfurt – Proceedings, ed. by Jörg Helbig, 47–64. Trier: WVT.

Busse, Beatrix. 2010b. “Recent Trends in New Historical Stylistics”. Language and Style. In Honour of Mick Short, ed. by Dan McIntyre & Beatrix Busse, 32–54. London: Palgrave Macmillan.

Busse, Beatrix. 2012. “Historical Text Analysis: Underlying Parameters and Methodological Procedures”. Methods in Contemporary Linguistics, Trends in Linguistics Series, ed. by Andrea Ender Adrian Leemann & Bernhard Wälchli, 285–308. Berlin: Mouton.

Busse, Ulrich & Beatrix Busse. 2010. “Shakespeare” Historical Pragmatics, ed. by Andreas H. Jucker & Irma Taavitsainen, 247–281. Berlin: Mouton.

Carew, Richard. 1595/96. “The Excellency of the English Tongue”. Linguistics in Great Britain, vol. I, History of Linguistics (1971), ed. by Wolfgang Kühlwein, 38–42. Tübingen: Niemeyer.

Gazzard, Hugh. 2009. “An Act to Restrain Abuses of Players (1606)”. The Review of English Studies 61(251): 495–528.

Jucker, Andreas H. 1997. “The Discourse Marker well in the History of English”. English Language and Linguistics 1: 91–110.

OED = Oxford English Dictionary online. 5 Jun. 2012. http://www.oed.com/.

Taavitsainen, Irma & Susan Fitzmaurice. 2007. “Historical Pragmatics: What it is and how to do it”. Methods in Historical Pragmatics, ed. by Susan Fitzmaurice & Irma Taavitsainen, 11–36. Berlin: Mouton.

Appendix 1: List of Shakespeare's works referred to in this paper

Appendix 2: List of modern editions of Shakespeare's works (included in Vol. IX Substantive Variants (1980) of Spevack (1968–1980)

ALEX = Alexander, Peter, ed. Works. London (1951).

ARD2 = The [New] Arden Shakespeare. Ellis-Fermor, Una et al., eds. Works. London (1951–).

CAM2 = The second Cambridge Shakespeare. Wright, William Aldis, ed. Works. 2nd ed. 9 vols. London (1891–93).

CAM3 = The New Cambridge Shakespeare. Wilson, John Dover et al., eds. Works. 39 vols. London (1921–66).

EVNS = Evans, G. Blakemore et al., eds. The Riverside Shakespeare. Boston (1974).

KIT1 = Kittredge, George Lyman, ed. Works. Boston (1936).

PEL2   = The complete Pelican Shakespeare. Harbage, Alfred, gen. ed. Complete Works. Baltimore (1969).

SIS = Sisson, Charles Jasper, ed. Complete Works. London (1954).