This article is based on a paper presented at the Directions in English Language Studies (DELS) conference, April 6-8, 2006 in Manchester. I would like to thank Ulrich Busse, Alexander Brock, Christian Mair, Ingo Plag, Naomi Hallan, Dietmar Schneider, and two anonymous reviewers for useful comments on earlier drafts of this article. I remain solely responsible for all remaining shortcomings. AAA - Arbeiten aus Anglistik und Amerikanistik Band 33 (2008) Heft 1 Gunter Narr Verlag Tübingen Investigating the Morphological Productivity of Verbal Prefixation in the History of English Anne Schröder The question of productivity has become one of the central empirical problems in English word-formation. In current approaches to morphological productivity, the existence of a continuum is acknowledged, at one end of which we find completely unproductive patterns and at the other highly productive ones, with a number of intermediate cases in between. As a consequence, various ways of measuring how productive a particular pattern is have been proposed in the literature. This paper highlights results from an ongoing study of verbal prefixation in English, of the kind download, upgrade, overachieve, underexpose, etc., in which different approaches to measuring productivity have been (or will be) applied. One of these - a dictionary-based investigation carried out on the basis of the OED - will be presented in detail and backed up with results from corpus-based research. The results of the present study show that the verbal patterns under investigation behave similarly to a number of other derivational patterns and methods of enlarging the vocabulary and that verbal prefixation of the type investigated has never ceased to be a productive means of verbal wordformation in English. 1. Introduction A number of authors have recently noted that the question of productivity has become one of the central empirical problems of word-formation (cf. e.g. Anne Schröder 48 1 Probably the most exhaustive list of definitions of productivity is provided by Rainer (1987: 188-190); this is also taken up by Bauer (2001: 25). 2 For a critical discussion and evaluation of the various measurements of productivity, see e.g. Schröder (2007). 3 See also Scheible (2005). 4 The examples given by Kastovsky are: outbid, outlive, outride and override, overreach, oversleep and underbid, undervalue, underrate. Bauer 2003: 70; Plag 1999: 1). However, when investigating the productivity of a particular word-formation pattern, one quickly realises that a number of rather different definitions of productivity are given in the literature 1 and that, as a consequence, various ways of measuring the productivity of a particular pattern have been proposed. In most recent publications, authors agree on three appropriate ways of measuring productivity: 1. measurements based on dictionary listings 2. measurements based on the analysis of corpora 3. measurements based on psychological tests of native speakers’ intuition, i.e. elicitation tests (cf. e.g. Bolozky 1999). 2 For the present study, which aims to investigate the productivity of verbal prefixation in the history of the English language, a definition of productivity based on the number of new forms occurring in a specified period of time appears most useful. According to Rainer (1987: 190), this measure has been most accurately formulated by Neuhaus (1973). This definition approaches productivity from a diachronic perspective, it is quantitative in nature and draws upon actual words. It can be investigated on the basis of a reliable chronological dictionary such as the OED (cf. e.g. Neuhaus 1973) and thus represents an application of the first of the three measuring types proposed. The verbs investigated are of the kind exemplified by download, upgrade, overachieve and underperform. These four verbs are recent creations: download is first attested in 1980, underperform in 1976, upgrade in 1920, 3 and overachieve in 1953 (OED 2002) and the question which I shall try to answer is whether these creations indicate an increase in productivity of a word-formation pattern which is often regarded as obsolescent or even fossilized, i.e. unproductive. Kastovsky, for instance, believes such formations to be moderately productive with the metaphorical meaning ‘to do sth. in excess’ or ‘below the expected limit’, 4 but unproductive in their literal meanings, as the latter are said to have been replaced by phrasal verbs in the Middle English period (2006: 254). Investigating the Morphological Productivity of Verbal Prefixation 49 5 See also Schäfer (1989) for corrections and antedatings of the OED. 6 Because of the possibility of analogical coining, which is very apparent in pairs such as e.g. over-achieve - under-achieve and upgrade - downgrade, pairs of prefixes were chosen. 7 The morphological structure of this complex lexeme is difficult to analyse. The seemingly obvious analysis as [in [[crass] [ate]]], defining crass as the root, and crassate as the base to which the prefix inis attached to form the final complex lexeme, is contradicted by the fact that the OED does not attest the existence of the lexeme *crassate. A lexeme *incrass is not attested either. I do not believe that an analysis involving etymological information (L. incrassat-, ppl. stem of incrassare > E. incrassate) is available to the majority of English speakers, which however does not render this complex word unanalysable, as the adjective crass does exist. I therefore assume a non-hierarchical structure, such as: [[in] [crass] [ate]]. 8 The etymology is not entirely indisputable as both noun and verb are first attested in 1895; the latter, however, only in a dictionary, the first textual evidence dating from 1951 (OED 2002). 2. Data sampling and results The present investigation is based on the 2002 electronic version of the OED on CD-ROM. The advantage of this historical dictionary as a source of data lies in its attempt to cover the whole extant word stock of modern English, from their earliest records up to the present day and including all registers (OED 2002: preface). In addition, the OED provides the first citation dates of words, which “can be taken as a rough guide to the age of a word, even though it is clear that a word may have been in spoken use long before it was written down for the first time, and the first quotation in the OED may not be the very first time it was written down” (Jucker 1994: 150). 5 The following pairs of prefixes 6 were investigated: in-, out-, up-, down-, over-, under-, onand off-. As will become evident, the term PREFIX VERB is used in the present study as a cover term for verb-formations based on a number of morphological rules, whose common elements are: a) their first part, i.e. any of the eight prefixes mentioned above; b) the word class of the resulting derived complex lexeme, which must be a verb. The element following any of the eight prefixes may be a simple verb, e.g. to out-act, or an adjective, e.g. to out-active, a noun, e.g. to out-admiral, or a proper name, e.g. to out-Herod, (all free lexemes), or else of a bound (or free) complex lexeme, such as adjective + suffix, as in e.g. to incrassate. 7 Furthermore, many of the verbs subsumed under the term PREFIX VERB in the following analyses, such as downgrade, could be analysed as results of an inversion process from phrasal combinations (cf. Plag 2003: 143). These are included, just as are verbs derived by backformation, as e.g. to over-house from over-housed, or by conversion, e.g. to offprint. 8 This study is thus clearly output-oriented, as in word-formation we generally deal with heterogeneous patterns, consisting of a number of sub-patterns, which yield the same Anne Schröder 50 9 See also Börger, who, after a thorough discussion of the status of OVER+X lexemes states that these constitute “borderline cases situated somewhere between prefixation and compounding” (2006: 16). 10 Dalton-Puffer and Plag (2001: 242), referring back to Rainer (1993), also judge this semantic criterion to be “the most important criterion to distinguish a compound element from a suffix”. 11 Downis classified under the entry down, adv. and referred to as an adverb in the subentry ‘downin combination’. Overis not labelled in the entry’s head but is referred to as a prefix in the descriptive paragraph on the entry. The examples given, however, are referred to as compound verbs. Outis referred to as both prefix and adverb in the entry for out-. (OED 2002) 12 In the context of this paper, a word is considered to be a neologism at the time of its earliest citation in the OED. For a more detailed discussion of the term ‘neologism’ and its various definitions and applications, see Elsen (2004: 19-23), Busse (1996), Hohenhaus (1996: 15-68), and Raab-Fischer (1994: 3-6). outputs. Very frequently it is difficult, if not impossible, to know with certainty which morphological rule accounts for a specific output. I am aware of the fact that the term PREFIX in the present context is - and has been - problematic. Many of the verbs investigated here have actually been described as compound verbs (cf. e.g. Marchand 1969: 96) or as prefix compounds, “suggesting a status somewhat ambiguous between true compounds and some other kind of derivational formation” (Lass 1994: 198). 9 But, as Adams (2001: 71) points out, the initial elements of the verbs investigated (which she names particles) have much in common with prefixes, especially when they are semantically distinct and are used to form series of verbs, such as out-, over-, under-. 10 This view is partly supported by, for example, their classification in the OED. As mentioned above, a number of verbs relevant to my investigation are given under the articles on the respective prefixes in the OED and at least up-, in-, under-, on-, offare classified as such in the OED, although over-, down-, outare not. 11 Moreover, there seems to be little difference between underas in underlet (in the sense ‘to let to a sub-tenant’) and subin the synonymous sublet. The productivity of these eight prefixes was investigated as follows: I looked for verbs prefixed by any of the eight prefixes in the alphabetical list of main entries in the OED and noted the date of their earliest citation, which can be taken to correspond approximately to the date of their coinage. In addition, verbs which were not listed as main entries but were given under the main entry of a related headword (usually the related noun or adjective), as well as verbs given in the articles for the respective prefixes, were also noted with their earliest citation dates. On the basis of this search I was able to illustrate the developments or changes of productivity of the eight verbal prefixes in question, by analysing the absolute number of neologisms 12 created in time spans of 50 years, from the earliest periods attested in the OED to the present, resulting in graphs such as the following: Investigating the Morphological Productivity of Verbal Prefixation 51 13 See also Scheible (2005) on the present-day productivity of verbal upand down-. 14 See Table 4 in the Appendix, where the relevant figures are put into italics and set in bold type. Figure 1: OED analysis for verbs with the prefix down- 13 Figure 2: OED analysis for verbs with the prefix up- A separate graph was generated for each of the eight prefixes (see also Figure 8 to Figure 13 in the Appendix). Combining all eight graphs into one diagram gives a rather blurred picture, as the prefixes for which I have a comparatively low total number of neologisms (as e.g. verbs with the prefix down-, cf. Figure 1) are not well represented. In order to gain a clearer picture, the changing patterns of productivity were compared in terms of the percentages of contribution, during the various periods, to the total number of neologisms for each pattern, as illustrated in the following figure: 14 Anne Schröder 52 15 For instance, Booij (1977: 5) believes productivity to be a matter of ‘all or nothing’. But it is primarily in inflectional morphology that productivity can be easily regarded as ‘all-or-none’ (Clark 1993: 127). Therefore many scholars take productivity as a matter of degree and believe some morphological processes to be more or less productive than other morphological processes and thus consider productivity to be a quantitative concept. See also Kettemann (1988: 13). Figure 3: OED analysis for verbs with the prefixes in- , out-, under-, over-, on-, off-, up, down-, in percentages In this diagram, increases of productivity during the Middle English and late Middle English period (although at slightly different points in time for the different prefixes) are clearly visible, as is a peak of productivity for nearly all patterns during the 16 th and 17 th centuries. Similarly, all patterns show a third peak of productivity during the 19 th century. If productivity is taken as qualitative concept, i.e. a morphological process is defined as either productive or unproductive, 15 all patterns apart from offcan be considered as currently productive, as they have produced at least one neologism since 1950. Table 4 in the Appendix presents the absolute numbers of neologisms created during each time span, together with percentages relative to the total number of formations during the respective time spans and relative to the total number of neologisms created by each word-formation patterns. Idealising the development of productivity of all prefix verbs, taking the total number of neologisms created in each of the periods calculated as a percentage of all neologisms (see last column of Table 4 in the Appendix), gives the following simplified graph: Investigating the Morphological Productivity of Verbal Prefixation 53 Figure 4: OED analysis of patterns of productivity of English verbal prefixation, in percentages To many readers, the course of the graph in Figure 4 will look remarkably familiar, as it strongly resembles a graph by Wermser (1976: 24), which has been used in a number of reference works on the lexical development of the English language, for example in Crystal’s The Cambridge Encyclopedia of the English Language (2003: 72) or Nevalainen’s account on “Early Modern English Lexis and Semantics” in the 3 rd volume of The Cambridge History of the English Language (1999: 339). The striking correspondences between the graph in Figure 4 and Wermser’s graph become evident in Figure 5. Although there are some differences in methodology (Wermser counted the neologisms in time spans of 20 years and based his study on the shorter version of the OED), the same peaks of productivity are visible in both studies. Anne Schröder 54 16 As demonstrated by Durkin (2002), the documentation in the later editions of the OED has been improved immensely and the graphs presented here are thus surely more accurate than the one given by Wermser. Figure 5: Patterns of productivity of prefix verbs in relation to the lexical summits of English However, although it has been widely accepted that the number of first citations in the OED basically reflects the lexical growth of the English language, which can be represented graphically, as here or in Wermser (1976), the accuracy of this method has been equally widely criticised (e.g. Brewer 2000; Kastovsky 2006: 266/ 267; Schäfer 1980; Willinsky 1994). Most points of criticism are summarised by Algeo: The neat and impressive-looking line graphs that have been drawn to show the peaking of word-making in the vigorous, language intoxicated high Renaissance, its deep valley of decline in the eighteenth century, and its subsequent rise to a new, if lesser, high in the mid-nineteenth century show nothing about the language. What they show is the extent and assiduousness with which the OED volunteers read and excerpted books. Shakespeare was overread; the eighteenth century under-read - that is what the graphs show. (Algeo 1998: 64) Although the criticism thus voiced might be somewhat exaggerated and not entirely justified, especially for newer editions of the OED, 16 any such repre- Investigating the Morphological Productivity of Verbal Prefixation 55 17 It is far beyond the scope of the present study to perform a manual analysis for each of the eight prefixes, so the analysis has been limited to one pair of antonymic prefixes, namely underand over-. 18 For a detailed description of the Lampeter Corpus, see Claridge (2000: chapter 2) or Schmied (1994). sentation, whether Wermser’s or mine, may yet provide only an imprecise picture of the growth of the English lexicon as a whole. However, it is plausible to assume that any distortions would affect more or less all word-formation processes to the same extent. I therefore feel justified in claiming that the diachronic development and increases and decreases of productivity of the eight prefixes investigated here have not taken a different course from that of other word-formation processes and methods of enlarging vocabulary, investigated by similar methods. These claims can be substantiated by corpus-based measurements. If, for example, the number of types, tokens and hapaxes obtained for prefix verbs with underand over- 17 in the Lampeter Corpus of Early Modern English Tracts, a collection of prose writings from 1640 to 1740 (Claridge 2000: 4), 18 are compared to those obtained from the Freiburg update of the Lancaster- Oslo/ Bergen Corpus (FLOB), a corpus which represents the state of British English in 1991 (see e.g. Mukherjee 2002: 30), we get the following results: Table 1: Number of types, tokens and hapaxes for prefix verbs with the prefixes underand overin the Lampeter Corpus and the FLOB Corpus overunder- Lampeter types 40 10 tokens 172 478 hapax legomena 19 3 FLOB types 42 21 tokens 169 413 hapax legomena 20 7 Table 1 shows that with regard to the number of types and tokens for the prefix verbs in overand undersome diachronic changes can be observed. While the number of types of prefix verbs in overhas remained comparatively stable, the number of types of prefix verbs in underhas doubled in the FLOB Corpus. However, the number of tokens of prefix verbs in underhas slightly decreased in PDE, while that for prefix verbs in overhas not changed diachronically. With regard to the hapaxes for these two types of verbal prefixation, Table 1 illustrates that the number for overhas remained relatively stable, while that for underhas more than doubled. Since the Anne Schröder 56 19 Börger does not define class D. In the context of the present analysis, semantically opaque, strongly lexicalised and frequent words, such as understand were classified in class D. number of hapaxes, i.e. those words occurring only once in a given corpus, are especially suited as an indication of morphological productivity (see e.g. Bauer 2004: 51), the fact that their numbers remained stable for the prefix overand more than doubled for underalso suggests that verbal prefixation as a word-formation process has not weakened diachronically. In addition, Kastovsky’s claim that prefix verbs are moderately productive with the metaphorical meanings of ‘to do sth. in excess’ or ‘below the expected limit’ only (2006: 254; see also above, p.2), also needs to be qualified. With Marchand, one can distinguish up to six semantic groups for the prefix overand five for the prefix under- (Marchand 1969: 97-100), and his scheme has been simplified and successfully applied to corpus data of Present-day English by Börger (2006), who distinguishes the following four semantic classes: A. negatively conceived excess (or deficiency) in terms of scalarity (e.g. to overreact); B. spatial meaning, literally or figuratively related to the concrete spatial prepositional meaning (e.g. to overshadow); C. abstract meaning conceptualised in terms of scalarity, connected indirectly to spatial meaning, without negative connotations (e.g. to overtrump); D. other. 19 If Börger’s classification scheme is applied to the prefix verbs found in the Lampeter Corpus and the FLOB Corpus, the distribution of the prefix verbs with overand underis as follows: Table 2: Type, token and hapaxes distribution of overand underacross semantic classes in the Lampeter Corpus prefix sense A B C D total overtypes 14 20 6 - 40 tokens 33 125 14 - 172 hapaxes 9 7 3 - 19 undertypes 2 5 1 2 10 tokens 6 53 19 400 478 hapaxes 1 2 - - 3 Investigating the Morphological Productivity of Verbal Prefixation 57 20 As Brinton and Traugott (2005: 127) point out, prepositional rather than phrasal verbs represent the functional replacement for prefixed verbs. 21 A happy exception is Mair, who states that “verbal prefixation, though increasingly marginalized, was never quite dead” (2006: 63). Table 3: Type, token and hapaxes distribution of overand underacross semantic classes in the FLOB Corpus prefix sense A B C D total overtypes 24 13 3 2 42 tokens 37 109 3 20 169 hapaxes 15 2 3 - 20 undertypes 9 9 1 2 21 tokens 30 98 1 284 413 hapaxes 4 2 1 - 7 This semantic analysis of the prefix verbs with underand overshows that - although the metaphorical meaning of ‘excess’ (sense A) has indeed gained importance diachronically - the importance of sense B, which is more clearly still related to the original spatial meaning, has equally increased for the prefix underand that other senses are also still productive in Presentday English. However, a caveat is necessary: I am aware of the fact that the two corpora differ in structure and that a comparison of data obtained from these two corpora can only allow for some very cautious assumptions as, to some degree, I am comparing “apples and oranges” (Lindquist and Levin 2000). In addition, only two of the 8 prefixes discussed were taken as test cases. Nevertheless, I believe that the comparison has some value and produces some worthwhile results. 3. Prefix verbs from a diachronic perspective Both the results from two corpora and the investigation carried out on the basis of the OED do not suggest a complete abandonment of this derivational process, whether through the introduction of postpositional devices in phrasal and prepositional verbs 20 or through the massive influx of French and Latin borrowings and loan-translations, as claimed in a number of reference works (e.g. Burnley 1992; Kastovsky 1992; Kastovsky 2006). 21 Anne Schröder 58 22 This phenomenon is usually only described by quasi-synonyms such as for instance rise - mount - ascend or ask - question - interrogate (cf. e.g. Kastovsky 2006: 248). One explanation for the results presented might be that what I here term ‘prefixation’ may also be referred to as ‘compounding’ (see above, p.2). The prefixes described here are different from many of those, for instance, in the study by Hiltunen (1983; cf. also Kastovsky 2006: 235-237). It may be a crucial point of the investigation that the elements which can synchronically be justifiably classified as prefixes may diachronically have arisen as compounds. The synchronic delimitation between compounding and affixation with these constructions is not always straightforward, as constituents of compounds can change into affixes (Kastovsky 2006: 212). But, whatever their status diachronically, the results here clearly illustrate that in the course of history, the type of verb-formation under discussion has never become unproductive. A reason for this continued productivity may be the structural and stylistic versatility of prefix verbs. As illustrated in Figure 6, the word-formation potential of the forms investigated is greater than the one of phrasal verb constructions: Figure 6: Structural versatility of prefix verbs as compared to verb-particle constructions phrasal verb prefix verb to go under to under-perform * the going-under enterprise the under-performing work force * a goer under an under-performer ? the going under of the under-performance of Thus, while adjectives and nouns are derived easily from prefix verbs, this proves much more difficult, if not impossible, with phrasal verbs. In addition, as illustrated in Figure 7, in a number of ways these constructions hold an intermediate position between analytic, native, stylistically informal phrasal constructions on the one hand and synthetic, foreign, stylistically formal Latinate or Romance borrowings on the other, with non-Germanic prefixation falling between the latter and the native prefix verbs. Thus, prefix verbs also contribute to the stratification of the English lexicon and the complexity of lexical fields. 22 Investigating the Morphological Productivity of Verbal Prefixation 59 23 Dates in brackets indicate the dates of the first attestation. 24 Outlet and underbear are considered rare in Present-Day English. 25 Outgo in the sense of ‘to outstrip in going’ is, of course, not obsolete in Present-Day English. 26 Underbear is actually first attested earlier (c950), but only in the sense of ‘to sustain’. With the sense of ‘to support’ and ‘to bear up’ it is first attested in 1382. Figure 7: Stratification of the lexicon 23 phrasal verb prefix verb (native) prefix verb (Latinate) borrowing to bear up to underbear (1382) to support (1382) to go out to outgo (†) (c825-1905) to exit (1607) to go in to ingo (†) (c900-1382) to enter (a1300) to let out to outlet (1592) to egress (1578) to set off to offset (1792) to counterbalance (1611) to print off to offprint (1895) to deprint (1909) to underlet (1819) to sublet (1766) Although some of the prefix verbs in Figure 7 are obsolete in Present-Day English (e.g. to outgo and to ingo) 24 , they have all coexisted with synonymous borrowings, sometimes for a substantial period of time. Thus outgo 25 was used side by side with exit for 300 years, and seems to have been firmly established in the English lexicon by the time the latter arrived. The same probably holds for ingo and the synonymous loanword enter. Some prefix verbs were coined simultaneously with the arrival of the corresponding Latinate borrowings, as in the case of underbear and support, both of which are first attested in 1382 in the Wycliffe Bible. 26 Others, such as to outlet and to offset, may even have been created after the arrival of the corresponding loanword. Anne Schröder 60 27 See Jones (1953: 94ff.) on the puristic movement during the Early Modern English period. 28 However, as Brinton (1988), Hiltunen (1983), and Lutz herself (1987) demonstrate, many OE prefixes nevertheless became obsolete. 29 However, see also Algeo’s comment on p.2 above. Loan-translations, such as for instance outgo imitating Latin exire (OED), have been an accepted (native) means of extending the vocabulary since the Old English period (cf. Kastovsky 1992: 300ff.; Kastovsky 2006: 217) and they were especially frequent during the Middle English period (Marchand 1969: 99). Although Hiltunen is certainly correct in pointing out that “there are so many factors involved in the study of loan-syntax that even in the case of positive correspondence it is hard to tell where they are due to the direct influence of the Latin, and where OE [Old English, A.S.] and Latin constructions just happen to overlap” (Hiltunen 1983: 139), I believe that this Latinizing practice strengthened the prefixes investigated here. They could be seen as a native alternative to foreign (and frequently opaque) loanwords, and this is why they were also resorted to by linguistic purists in the 16 th and 17 th centuries as a method of increasing lexical resources (Nevalainen 1999: 360). 27 In addition, as Lutz notes, prefix formations - irrespective of their etymological origin - also correspond to the morphological and phonological structure characteristic of the more formal registers of the English language since the Later Middle Ages. She mentions polysyllabic structures with an unaccented first element, which the rhythmical pattern of the post-specifying phrasal verbs does not produce (Lutz 1997: 285/ 286). Thus, she is presumably correct in believing that “the contribution of the Norman Conquest to the structural make-up of the Modern English lexicon is more of a preserving nature, as it has in fact stabilized prefixation as a means of verb formation for the more formal registers of English” (ibid.: 287). The evidence from this dictionary-based account confirms Lutz’s view that “[s]ome native prefixes have continued to be productive until today, in spite of massive French and Latin influence; they are combined with both native and borrowed stems” (ibid.: 260) and that “prefixing has not been abandoned altogether in English” (ibid.: 284). 28 The fact that prefix verbs, unlike phrasal verbs and other multi-word verbs, are not considered colloquial or informal and can be used in academic writing might also explain the last increase of productivity, at the end of the 19 th century, as shown in the graph. 29 Increased linguistic awareness and subsequent prescriptive tendencies during that period seem to have played a negative role in the use of phrasal verbs in writing (Claridge 2000: 199), and favoured the use of Latinate loanwords and loan-translations. This might also explain why most prefixes exhibit a recent decrease of productivity (cf. Investigating the Morphological Productivity of Verbal Prefixation 61 30 Cf. Claridge (2000: 103ff.) for a discussion. As Claridge makes clear, multi-word verbs can no longer be considered unsuitable for the more formal written registers of language: these constructions seem to be used at all levels of language in Present-Day English. However, see also Biber et al. (1999: 408-412) on the register distribution of phrasal verbs in Present- Day English. According to Biber et al. (1999: 415-422), there seems to be a difference between phrasal verbs and prepositional verbs with regard to register distribution, the latter being more easily accepted in academic prose than the former, and apparently having fewer informal overtones (ibid.: 415). 31 Cf. also the ‘download and to down’load, according to the OED. 32 See also Plag (2006). Figure 3 above), as the use of phrasal constructions seems to have become acceptable even in more formal registers of language. 30 Of course a full investigation of the present-day productivity of the prefix verbs requires the application of the other productivity measurements mentioned at the beginning of this article. Such investigations are not within the scope of this paper but I will briefly refer to some other studies, which are of interest in the present context. In Schröder (2007), for example, corpusbased measures are applied to investigate the productivity of prefix verbs with overand underin comparison with corresponding particle verbs. It is shown that particle verbs are not, generally speaking, more productive than the corresponding prefix verbs, as in this investigation it is only the particle over which is more productive than the verb formations with the prefix over-. The particle under is only more productive with regard to some of the measurements applied. In Scheible (2005) several measurements of productivity are applied to illustrate that prefix verbs with upand downare of moderate productivity in Present-Day English and that “the existence of a synonymous phrasal verb will generally block the coining of the corresponding” prefix form (Scheible 2005: 192), so that possible limitations of productivity as well as the relationship of stress alternations need to be taken into account. Hence the lack or the arbitrariness of stress alternation especially with later formations, e.g. to ‘output (1858) vs. the ‘output as opposed to the earlier (and possibly obsolete) to out’put (c. 1300), 31 should be systematically investigated. However, given that Plag et al. (2006) convincingly showed that a considerable amount of stress assignment in noun-noun compounds is unpredictable and that none of the commonly held hypotheses about compound stress assignment makes valid predications, similar facts may account for the stress assignment in the constructions under investigation here. 32 4. Conclusion Although the present-day productivity of prefix verbs needs to be investigated with additional measurements, on the basis of the present investigation alone Anne Schröder 62 a number of generalisations with regard to the diachronic productivity can already be made. The widespread belief that verbal prefixation weakened from the Middle English period onwards because of its gradual replacement, either by the postposition of the prefixes to form phrasal verbs, or by the adoption of Romance and Latinate loanwords, clearly needs to be reconsidered. The results of the present study show that at least the verbal patterns under investigation behave similarly to a number of other derivational patterns and methods of enlarging the vocabulary. Indeed, this type of verbal prefixation has never ceased to be a productive means of verbal word-formation in English. At this point in the study, it seems reasonable to claim that prefix verbs, as investigated and described here, are structurally more versatile than phrasal verbs and stylistically, they occupy an intermediate position between phrasal verbs and prepositional verbs on the one hand and Latinate/ Romance loan words on the other. Their versatility seems to have secured the position of this word-formation process up to the present-day English period. This assumption will clearly need to be qualified in the subsequent stages of the study, on the basis of corpus data, by taking register variation into account. A qualitative investigation, either on the basis of corpus data and/ or on the basis of elicitation tests, will consider the stylistic stratification indicated in Figure 7 and the importance of analogical coining and backformation. Nevertheless, the results of a historical survey based on dictionary records and corpus findings can already provide valuable information about the extent and persistence of this important mechanism for vocabulary extension. References Primary sources The Oxford English Dictionary (2 nd Edition) on CD-ROM, Version 3.0 (2002). Oxford: OUP. The Lampeter Corpus of Early Modern English Tracts (1999). (ICAME Collection of English Language Corpora, 2 nd edition.) The Freiburg-LOB Corpus of British English (1999). (ICAME Collection of English Language Corpora, 2 nd edition.) Secondary sources Adams, Valerie (2001). Complex Words in English. Harlow, London: Longman. Algeo, John (1998). “Vocabulary.” In: Suzanne Romaine (ed.). The Cambridge History of the English Language, Vol. IV, 1776-1997. Cambridge: CUP. 57-91. Bauer, Laurie (2001). Morphological Productivity. Cambridge: CUP. Bauer, Laurie (2003). Introducing Linguistic Morphology (2 nd Edition). Washington: Georgetown University Press. Bauer, Laurie (2004). A Glossary of Morphology. Edinburgh: Edinburgh University Press. Investigating the Morphological Productivity of Verbal Prefixation 63 Biber, Douglas et al. (1999). The Longman Grammar of Spoken and Written English. Harlow: Pearson Education. Börger, Claudia (2006). Word-formation Processes from a Cognitive Perspective. An Analysis of Complex Prepositional Lexemes. PhD thesis, Englisches Seminar, Leibniz- Universität Hannover. Bolozky, Shmuel (1999). Measuring Productivity in Word-formation. The Case of Israeli Hebrew. Leiden: Brill. Booij, Geert (1977). Dutch Morphology. A Study of Word Formation and Generative Grammar. Lisse: The Peter de Ridder Press. Brewer, Charlotte (2000). “OED Sources.” In: Linda Mugglestone (ed.). Lexicography and the OED. Pioneers in the Untrodden Forest. Oxford: OUP. 40-58. Brinton, Laurel (1988). The Development of English Aspectual Systems. Aspectualizers and Post-verbal Particles. Cambridge: CUP. Brinton, Laurel and Elizabeth Closs Traugott (2005). Lexicalization and Language Change. Cambridge: CUP. Burnley, David (1992). “Lexis and Semantics.” In: Norman Blake (ed.). The Cambridge History of the English Language, Vol. II, 1066-1476. Cambridge: CUP. 409-499. Busse, Ulrich (1996). “Neologismen. Der Versuch einer Begriffsbestimmung.” In: Martin Gellerstam et al. (eds.). Euralex ‘96 Proceedings I-II. Papers Submitted to the Seventh EURALEX International Congress on Lexicography in Gothemburg, Sweden. Göteborg: Göteborg University. 645-658. Claridge, Claudia (2000). Multi-word Verbs in Early Modern English. A Corpus-based Study. Amsterdam: Rodopi. Clark, Eve V. (1993). The Lexicon in Acquisition. Cambridge: CUP. Crystal, David (2003, 2 nd edition). The Cambridge Encyclopedia of the English Language. Cambridge: CUP. Dalton-Puffer, Christiane and Ingo Plag (2001). “Categorywise, Some Compound-type Morphemes Seem to Be Rather Suffix-like: On the Status of -ful, -type, and -wise in Present-day English.” Folia Linguistica 34/ 3-4. 225-244. Durkin, Philip (2002). “Changing Documentation in the Third Edition of the ‘Oxford English Dictionary’. Sixteenth-century Vocabulary as a Test Case.” In: Teresa Fango / Belén Méndez-Naya / Elena Seoane (eds.). Sounds, Words, Texts and Change. Amsterdam: Benjamins. 65-82. Elsen, Hilke (2004). Neologismen. Formen und Funktionen neuer Wörter in verschiedenen Varietäten des Deutschen. Tübingen: Narr. Hiltunen, Risto. (1983). The Decline of the Prefixes and the Beginnings of the of the English Phrasal Verb. The Evidence from Some Old and Early Middle English Texts. Turku: Turun Yliopisto. Hohenhaus, Peter (1996). Ad-hoc-Wortbildung. Terminologie, Typologie und Theorie kreativer Wortbildung im Englischen. Frankfurt: Lang. Jones, Richard Foster (1953). The Triumph of the English Language. A Survey of Opinions Concerning the Vernacular from the Introduction of Printing to the Restoration. London: OUP. Jucker, Andreas (1994). “New Dimensions in Vocabulary Studies. Review Article of the Oxford English Dictionary, 2 nd edition on CD-ROM.” Literary and Linguistic Computing 9/ 2. 149-154. Kastovsky, Dieter (1992). “Semantics and Vocabulary.” In: Richard Hogg (ed.). The Cambridge History of the English Language, Vol. I, The Beginnings to 1066. Cambridge: CUP. 290-408. Anne Schröder 64 Kastovsky, Dieter (2006). “Vocabulary.” In: Richard Hogg / David Denison (eds.). A History of the English Language. Cambridge: CUP. 199-270. Kettemann, Bernhard (1988). Die Phonologie morphologischer Prozesse im amerikanischen English. Tübingen: Narr. Lass, Roger (1994). Old English. A Historical Linguistic Companion. Cambridge: CUP. Lindquist, Hans and Magnus Levin (2000). “Apples and Oranges: On Comparing Data from Different Corpora.” In: Christian Mair / Marianne Hundt (eds.). Corpus Linguistics and Linguistic Theory. Papers from the Twentieth International Conference on Computerized Corpora (ICAME 20), Freiburg i.Br. 1999. Amsterdam: Rodopi. 201-213. Lutz, Angelika (1997). “Sound Change, Word Formation and the Lexicon. The History of the English Prefix Verbs.” English Studies 78/ 3. 258-290. Mair, Christian (2006). Twentieth-Century English. Cambridge: CUP. Marchand, Hans (1969). The Categories and Types of Present-day English Word-formation. A Synchronic-diachronic Approach. München: Beck. Mukherjee, Joybrato (2002). Korpuslinguistik und Englischunterricht. Eine Einführung. Frankfurt: Lang. Neuhaus, Heinz (1973). “Zur Theorie der Produktivität von Wortbildungssystemen”. In: Abraham Ten Cate / Peter Jordens (eds.). Linguistische Perspektiven. Referate des VII. Linguistischen Kolloquiums in Nijmegen. Tübingen: Niemeyer. 302-317. Nevalainen, Terttu (1999). “Early Modern English Lexis and Semantics.” In: Roger Lass (ed.). The Cambridge History of the English Language, Vol. III., 1476-1776. Cambridge: CUP. 332-458. Plag, Ingo (1999). Morphological Productivity. Structural Constraints in English Derivation. Berlin: Mouton de Gruyter. Plag, Ingo (2003). Word-formation in English. Cambridge: CUP. Plag, Ingo (2006). “The variability of Compound Stress in English. Structural, Semantic and Analogical Factors.” English Language and Linguistics 10/ 1. 143-172. Plag, Ingo et al. (2006). “Variable Stress Assignment in Noun-noun Compounds. New Evidence from Corpora.” Paper presented at the DELS Conference, 6-8 April 2006, in Manchester. Raab-Fischer, Roswitha (1994). “A Hyperinflation of Lexical Mega-Monsters? Eine korpusgestützte Analyse zum Gebrauch der Wortbildungselemente ‘mega-‘, ultra-‘, ‘super-‘ und ‘hyper-‘ im heutigen Englisch.” Arbeiten aus Anglistik und Amerikanistik 19/ 1. 83-111. Rainer, Franz (1987). “Produktivitätsbegriffe in der Wortbildungstheorie.” In: Wolf Dietrich / Hans-Martin Gauger / Horst Geckeler (eds.). Grammatik und Wortbildung Romanischer Sprachen. Beiträge zum Deutschen Romanistentag in Siegen. Tübingen: Narr. 187-202. Rainer, Franz (1993). Spanische Wortbildungslehre. Tübingen: Niemeyer. Schäfer, Jürgen (1980). Documentation in the OED. Shakespeare and Nashe as Test Cases. Oxford: Clarendon. Schäfer, Jürgen (1989). Early Modern English Lexicography. Vol. II. Addition and Corrections to the OED. Oxford: Clarendon. Scheible, Silke (2005). “‘Upgrading’, ‘Downsizing’ and Co.: Revitalising a Moribund Wordformation Pattern in Present-day English? ” Arbeiten aus Anglistik und Amerikanistik 30/ 1-2. 177-200. Schmied, Josef (1994). “The Lampeter Corpus of Early Modern English Tracts.” In: Merja Kytö / Matti Rissanen / Susan Wright (eds.). Corpora Across the Centuries. Proceedings of the First International Colloquium on English Diachronic Corpora. Amsterdam: Rodopi. 81-89. Investigating the Morphological Productivity of Verbal Prefixation 65 Schröder, Anne (2007). “Theoretical and Empirical Approaches to Productivity. Evidence from English Verbal Morphology.” In: Sabine Volk-Birke / Julia Lippert (eds.). Anglistentag 2006 Halle. Conference Proceedings. Trier: WVT. 399-411. Wermser, Richard (1976). Statistische Studien zur Entwicklung des englischen Wortschatzes. Bern: Francke. Willinsky, John (1994). Empire of Words. The Reign of the OED. Princeton: Princeton University Press. Anne Schröder Martin-Luther-Universität Halle-Wittenberg Institut für Anglistik und Amerikanistik Anne Schröder 66 Figure 8: OED analysis for verbs with the prefix under- Figure 9: OED analysis for verbs with the prefix over- Figure 10: OED analysis for verbs with the prefix on- Appendix: Investigating the Morphological Productivity of Verbal Prefixation 67 Figure 11: OED analysis for verbs with the prefix off- Figure 12: OED analysis for verbs with the prefix out- Figure 13: OED analysis for verbs with the prefix in- Anne Schröder 68 Table 4: OED analysis for verbs with the prefixes in- , out-, under-, over-, on-, off-, up, downupdowninoutoverunderonoff- Total before 1000 1 0 6 1 39 8 7 9 71 0.51 0.00 0.81 0.17 5.23 3.29 21.88 11.54 2.67 1000- 1049 3 0 4 2 8 4 2 10 33 1.52 0.00 0.54 0.33 1.07 1.65 6.25 12.82 1.24 1050- 1099 0 0 0 0 1 0 0 0 1 0.00 0.00 0.00 0.00 0.13 0.00 0.00 0.00 0.04 1100- 1149 0 0 0 0 0 0 0 2 2 0.00 0.00 0.00 0.00 0.00 0.00 0.00 2.56 0.08 1150- 1199 0 0 0 1 2 1 0 1 5 0.00 0.00 0.00 0.17 0.27 0.41 0.00 1.28 0.19 1200- 1249 6 0 0 0 13 7 3 15 44 3.05 0.00 0.00 0.00 1.74 2.88 9.38 19.23 1.65 1250- 1299 17 4 12 14 12 1 3 8 71 8.63 13.33 1.62 2.33 1.61 0.41 9.38 10.26 2.67 1300- 1349 18 2 17 18 21 6 3 11 96 9.14 6.67 2.30 3.00 2.82 2.47 9.38 14.10 3.60 1350- 1399 7 0 39 16 38 29 0 4 133 3.55 0.00 5.28 2.67 5.10 11.93 0.00 5.13 4.99 1400- 1449 20 0 28 12 18 5 2 3 88 10.15 0.00 3.79 2.00 2.42 2.06 6.25 3.85 3.30 1450- 1499 2 0 32 9 8 0 0 0 51 1.02 0.00 4.33 1.50 1.07 0.00 0.00 0.00 1.91 1500- 1549 10 2 82 18 29 14 3 1 159 5.08 6.67 11.10 3.00 3.89 5.76 9.38 1.28 5.97 1550- 1599 35 1 134 80 116 21 0 2 389 17.77 3.33 18.13 13.33 15.57 8.64 0.00 2.56 14.60 1600- 1649 14 3 253 146 133 35 1 2 587 Investigating the Morphological Productivity of Verbal Prefixation 69 updowninoutoverunderonoff- Total 7.11 10.00 34.24 24.33 17.85 14.40 3.13 2.56 22.03 1650- 1699 3 1 54 62 50 19 1 0 190 1.52 3.33 7.31 10.33 6.71 7.82 3.13 0.00 7.13 1700- 1749 4 0 11 30 22 8 0 1 76 2.03 0.00 1.49 5.00 2.95 3.29 0.00 1.28 2.85 1750- 1799 3 1 8 28 18 8 0 1 67 1.52 3.33 1.08 4.67 2.42 3.29 0.00 1.28 2.52 1800- 1849 25 1 20 49 52 19 0 1 167 12.69 3.33 2.71 8.17 6.98 7.82 0.00 1.28 6.27 1850- 1899 17 2 33 75 100 26 5 6 264 8.63 6.67 4.47 12.50 13.42 10.70 15.63 7.69 9.91 1900- 1949 5 6 5 22 46 10 0 1 95 2.54 20.00 0.68 3.67 6.17 4.12 0.00 1.28 3.57 after 1950 7 7 1 17 19 22 2 0 75 3.55 23.33 0.14 2.83 2.55 9.05 6.25 0.00 2.82 Total 197 30 739 600 745 243 32 78 2664 (7.39) (1.13) (27.74) (22.52) (27.97) (9.12) (1.20) (2.93) (100) 100% 100% 100% 100% 100% 100% 100% 100% 100%