4.1 Outline

So far, the history of prefixation as a scaling device in English has received only scant attention. While verbal prefixes have been studied more widely, in particular with respect to the relation of prefixed and phrasal verbs (see Lutz Reference Lutz1997), literature on the diachrony of English nominal and adjectival prefixation in general is rather scarce, and those accounts which are available mainly provide inventories of prefixes, cataloguing their prototypical meanings without discussing issues of wider relevance (notable exceptions are Marchand Reference Marchand1969: 129–208 and Kastovsky Reference Kastovsky and Hogg1992: 377–81; see Dietz Reference Dietz2004: 562–4 for a survey on the literature and its criticism). Only recently has a fuller, more systematic description of Old English prefixes been attempted by Dietz (Reference Dietz2004).

Old English exhibits a greater degree of similarity to Present-day German than to Present-day English in that a larger number of booster prefixes is attested. In the following section, I will briefly introduce the items which function as booster prefixes in Old English. In a more detailed account, the study will then investigate the Old English roots of ME forsooth(e), an emphasizer which became very frequent at the beginning of the Middle English period. It will be suggested that forsooth – which is commonly considered to be a univerbated and lexicalized form of an Old English prepositional phrase for soþ ‘for truth’ – may alternatively be analysed as an (endingless) accusative singular neuter of an adjective forsoþ ‘very true’.

4.2 The inventory

In table 1 I list the items which have been suggested as booster prefixes modifying adjectives or adverbs in the research literature on Old English prefixes (Quirk & Wrenn Reference Quirk and Wrenn1957: 110; Kastovsky Reference Kastovsky and Hogg1992: 377–81, Reference Kastovsky, Hogg and Denison2006: 233–4, 263; Dietz Reference Dietz2004; Sauer Reference Sauer, Eitelmann and Stritzke2006; Mitchell & Robinson Reference Mitchell and Robinson2007: 58–9) and in the relevant Old English dictionaries (BT; Clark Hall 1960; the DOE for the letters A-F). Since the different dictionaries do not always agree on the classification of certain prefixes and on their acceptance of the relevant individual adjectives/adverbs as prefixations (not, for instance, as compounds or syntagms of intensifying adverb and adjective), I list all of the items in question alphabetically, together with all of their attested individual coinages; doubtful or ambiguous prefixes, most of which will be discussed later in the text, are marked in the table by? and their alternative interpretation is given in the last column. Single doubtful adjectives are also marked by ?, with explanations given in the respective footnotes.Footnote ¹²

Table 1.

Items suggested as booster prefixes modifying adjectives and adverbs

^a For a critical view on the relation of æl- and eall-, see Dietz (Reference Dietz2004: 587–8). For a more general account of the functions of eall/all in the history of English, see Buchstaller & Traugott (Reference Buchstaller and Traugott2006). OE eall-mihtig/æl-mihtig is not listed here, because it is generally considered to be a loan-translation of Latin omni-potens (see DOE s.v. eall-mihtig, æl-mihtig), which suggests its interpretation as an æl-compound.

^b For a discussion of borderline cases between compounds and prefixations (also labelled prefixoids), see Kastovsky (Reference Kastovsky and Hogg1992: 363–4).

^c I do not list unambiguous compound formations such as fela-feald ‘manifold’ or fela-sprecol ‘talkative, loquacious’ here (see DOE s.vv. fela-feald, fela-sprecol, fela-specol).

^d The DOE relates this word to the prefix for modifying the adjective scyldig or, alternatively, to a prefixed verb forscyldigian ‘only in past participle: made or become guilty/wicked’ (see DOE s.v. forscyldigian).

^e This could also be interpreted as a verbal prefix, i.e. the past participle of a verb for-swellan (which, however, is not attested in Old English; see DOE s.v. for-swollen).

^f Dietz (Reference Dietz2004: 602) points out that some of the coinages with fore- seem to be mergers of OE for- and fore-. Fore-mære and for-mære, for instance, are listed in one entry in the DOE (s.v. fore-mære/for-mære; cf. also the DOE s.v. fore-).

^g On forþ- as a nominal and verbal prefix, see Dietz (Reference Dietz2004: 601–2) and the DOE (s.v. forþ-).

^h For a full discussion of this prefix, see Dietz (Reference Dietz2004: 568–9) and, especially, Sauer (Reference Sauer, Eitelmann and Stritzke2006). Most of the coinages are modelled on Latin adjectives intensified by the Latin booster prefix prae-/pro- (cf. Latin prae-clarus – OE fræ-beorht ‘very bright’, Latin prae-pinguis – OE fræfætt ‘very fat’, etc.) and are therefore to be considered as loan formations (for the etymology of OE fræ-, see Sauer Reference Sauer, Eitelmann and Stritzke2006: 11–16).

ⁱ Verbs with the prefix ful- (cf. OE ful-don ‘achieve’, ful-endian ‘bring to an end’ or ful-fillen ‘fulfil’) have commonly been treated as ‘compound verbs’ (see, e.g., Kastovsky Reference Kastovsky and Hogg1992: 374–6). For their analysis as prefixed verbs, see Dietz (Reference Dietz2004: 584–95, 606). On the nominal forms, see Dietz (Reference Dietz2004: 595–6).

^j Because of its only sporadic occurrence Dietz (Reference Dietz2004: 591) now rejects the view of Holthausen, who regards in- as an intensifying prefix in in-frod (see Holthausen Reference Holthausen1934: s.v. in).

^k On the lexicalization of the prefix or-, see Dietz (Reference Dietz2004: 570).

According to the criterion that a productive prefix must have a high number of different types, regardless of the number of tokens (referred to as Reihenbildung ‘formation of series’ in Dietz Reference Dietz2004; for the relevance of the type–token relation, see section 4.4 below), a core of five unambiguous Old English booster prefixes emerges: for- (35 types),Footnote ¹³ofer- (12 types), frae-/fræ- (10 types), heah- (5 types) and or- (4 types).

For various reasons, most of the other linguistic elements listed above are not prototypical intensifying prefixes. Because they retain much of their original meaning, they are restricted to a limited set of derivational bases, and thus should be classified as compound adjectives. This suggests, for example, a purely qualitative interpretation of the prefixes ær- ‘. . . of old’ (e.g. ærglæd, ærgod ‘glorious/good of old’) and sin- (e.g. sinceald ‘perpetually cold’; see Dietz Reference Dietz2004: 593–4).

Other combinations, containing prefixes such as brego- or cyne-, are best interpreted as compounds with a metaphorically employed first element, i.e. ‘like a ruler/king’ (e.g. brego-rof ‘brave as a ruler’) or ‘royally, like a king’ (e.g. cyne-beald, cyne-god, cyne-ðrymlic ‘bold, excellent, glorious as a ruler’). A metaphorical interpretation is also most likely for deop-ðancol ‘deep-thinking > very thoughtful’, which implies the transfer that the better or more intense thoughts are to be found deep in one's mind.

Two of the prefixes – fela- and ful- – are semantically and/or syntactically polyfunctional, and their functions are not easy or (in the case of ful) even impossible to differentiate. Fela- ‘many’ (cf. DOE s.v. fela, MED s.vv. fēle indef. num., fēle adv., OED s.v. fele adv. (quasi-sb.) and a.) retains its numeral sense in forms such as felafeald ‘manifold’, felaspræce ‘talkative’ or felageonge ‘very much travelled’ (and other forms with verbal bases, none of which are listed above), but has a very clear intensifier meaning in felasynnig ‘very guilty’ or felamodig ‘very bold’.

Distinguishing between various functions proves almost impossible for combinations of ful + adjective. Since word division is not at all systematically coded or morphologically induced in Anglo-Saxon manuscripts, these may be analysed as combinations of the adverb ful plus an adjective or, alternatively, as prefixations. As a prefix, ful- may retain its original meaning ‘having no space empty, full’ with past-participle derivational bases in particular (e.g. full-ripod ‘fully grown’, full-mannod ‘fully peopled’ or ful-wepned ‘fully armed’). With most of the other forms listed above, it is almost impossible to decide whether we have an adverbial booster or a booster prefix. This is reflected in the treatment of these words in different dictionaries: while Clark Hall (1960), for instance, lists all of the items above as prefixations, the DOE only has two entries which are classified as prefixations, namely for full-cuð ‘well-known’ (22 occ.) and fullneah ‘very near’ (70 occ.). Campbell, on the other hand, clearly favours a prefix interpretation for most of the coinages, pointing to the fact that the ‘intensive prefix full- is subordinated in stress to the following element’ (Reference Campbell1959: §86). While there can be no doubt that ful was used as an adverbial booster in Old English (see Méndez-Naya Reference Méndez-Naya2003: 386–7), interpreting it as a prefix modifying adjectives and adverbs might be possible or even preferable for certain of the coinages listed above. This view is corroborated by the distributional patterns of types and tokens, since ful-, just like for-, has a number of rare or single occurrences (hapax legomena), but also some high-frequency lexemes, such as full-oft or full-neah (see below, section 4.4).

With regard to their source domains, Old English booster prefixes make use of the same domains which are used for the coinage of English adverbial boosters and Present-day German booster prefixes (see section 3 above and Peters Reference Peters1993). Even if the source domains of some of the coinages have become opaque due to lexicalization processes (e.g. in the prefixes for-, fræ-/frea- and or-), two source domains emerge unambiguously as the principal ones: (i) the local/dimensional domain and (ii) the quantitative domain.

1. (i)

2. (ii)

Footnote ¹⁴

The lack of expressions coined from category (iv) taboo/swear wordsFootnote ¹⁵ is most probably due to the nature of our extant Anglo-Saxon texts. Taboo booster prefixes are predominantly found in spoken language, particularly in colloquial language or slang. So it therefore comes as no surprise that they are not attested in our Old English written sources, most of which belong to a religious, highly literary or scientific genre.

4.3 OE for-

OE for- is clearly the booster prefix with the highest number of different types (35). In Old English, for- is a highly frequent and polyfunctional prefix which may be attached to nominal as well as verbal bases (just like Latin per-). In its much more frequent function as a verbal prefix,Footnote ¹⁶ it may code a negative meaning (e.g. forswerian ‘to commit perjury’, forwyrcan ‘to be guilty of’), but much more frequently expresses a perfective or telic aspect (e.g. forbærnan ‘to burn up, to burn to the end, to consume by fire’, fordon ‘to destroy’).

With regard to its use as a booster prefix, it is interesting to see that the etymologically related Latin per- is also used as a booster prefix for adjectives and adverbs, in particular in Medieval Latin (e.g. perdelectabilis ‘very enjoyable’, perhumilis ‘very humble’, peravidus ‘very greedy’, persaepe ‘very often’ etc.; see Stotz Reference Stotz2000: 132.1).

Its Old English booster use, however, is certainly not only due to Latin loan influence. The booster function is also attested in Old Norse (e.g. for-ríkr ‘very powerful’, for-vitr ‘very intelligent’) and, more importantly, for- can be shown to be a highly productive prefix in Old English. To test this issue, it is revealing to look at the number of DOEC occurrences of the individual adjectives and adverbs prefixed by for- (as given by the DOE):

This distribution of types and tokens is – according to Baayen (see, for example, Baayen Reference Baayen, Booij and Marle1993 or Baayen & Renouf Reference Baayen and Renouf1996) – tremendously important for the question of the productivity of an element. In essence, Baayen's research reveals that the number of words that occur only once in a given corpus (i.e. hapax legomena with respect to the given corpus) correlates with the number of neologisms and is therefore highly indicative of the productivity of a morphological element:

In view of these findings, for- can be considered a most productive prefix in Old English. As seen above, it is the booster prefix with the highest number of types (35). Moreover, in the large Old English corpus (c. 3 million Old English words are collected in the DOEC) we have more than 15 hapax legomena with for-,Footnote ¹⁷ but also a number of high-frequency words. Prototypically, the high-frequency items with more than 50 occurrences are used as adverbial downtoners (e.g. forneah, fornean) or boosters (e.g. forwel).

5.1 Introductory remarks

The above survey has shown that booster prefixes modifying adjectives and adverbs are much more frequent in Old English than in Present-day English. Among the core set of booster prefixes, for- is the most frequent prefix, with respect to both types and tokens. In the following section, I will suggest a further candidate for such a prefixation with for-, by trying to shed some light on the diachrony of ME/PDE forsooth ‘truly’. This analysis will also need to incorporate a discussion of the history of other Old English adverbials meaning ‘truly’, because forsooth(e) is merely attested six times in Old English and only becomes a high-frequency item in Middle English – first supplementing and eventually ousting the semantically and functionally similar adverbs and prepositional phrases soþlice, soþes and to soþe/to soþan (cf. Appenzeller-Gassmann Reference Appenzeller-Gassmann1961: 47–8; Swan Reference Swan1988: 259; Lenker Reference Lenker, Fischer, Rosenbach and Stein2003: 272–84, Reference Lenker, Fitzmaurice and Taavitsainen2007: 83–90).

5.2 Old English adverbials meaning ‘truly’

The core items signifying the concept ‘truth; true’ in Old English are the neuter noun soþ ‘truth’ and the formally identical adjective soþ ‘true’. The adverbial function – ‘truly’ – can be expressed by derived adverbs in -e or -es or by prepositional phrases. Derivational adverbs are soþe and soþlice (< soþlic adj.) in -e and soþes (with genitive suffix -es). Among the phrasal items, prepositional phrases with to (governing the dative) are most important, namely to soþe and to soþan, as opposed to the scarcity of for soþ and mid soþ. In the DOEC, soþlice (4,806 occ.) is by far the most frequent of these items, followed by the prepositional phrases to soþan (109 occ.), to soþe (66 occ.) and the derivation soþes (57 occ.). For soþ, by contrast, is attested only six times in Old English.

Apart from idiolectal differences,Footnote ¹⁸ the distribution of the various lexemes and prepositional phrases also shows a distinction between poetry and prose, as well as exhibiting functional differences in prose. The Old English poetical texts almost exclusively employ the phrase to soþe, most often in collocation with a verb of communication such as secgan ‘tell’:

1. (1)

   Gif þu him to soðe sægst hwylce þu selfa hæfst bisne on breostum.

   ‘If you tell him truly which exemplary precept you yourself hold in your bosom.’

   (GenA, B 570, DOEC)Footnote ¹⁹

2. (2)

   Secge ic þe to soðe, sunu Ecglafes, þæt . . .

   ‘I tell you for sure, son of Ecglaf, that . . .’

   (Beo 590, DOEC)

Soþlice, on the other hand, is extremely rare in poetry (25 instances in total), but is employed in various ‘layered’ functions in Old English prose: on the phrase level, it is – albeit rarely – used as a manner adjunct. Similar to the contexts cited above for to soðe, the adjunct use is predominantly found in direct speech with a first person subject in phrases such as ic secge soþlice ‘I tell (you) truly’.

1. (3)

   Nacode he scrydde, and swa ic soðlice secge, ealle nyd-behæfnysse he wæs dælende þam þe þæs behofodon;

   ‘The naked he clothed; and, as I truly tell, he distributed to every necessity of them that had need thereof.’

   (LS 8 (Eust) 8, DOEC)

These instances of soþlice (and similarly ME soothli and forsooth) as manner adjuncts with a verb of communication are crucial for all their additional uses as emphasizers, style disjuncts and pragmatic markers. First, this underlying and possible substitute phrase – ‘I tell you truly’ or ‘I tell you for sure’ – gives rise to the adverb's use as an intensifier or ‘emphasizer’ (Quirk et al. Reference Quirk, Greenbaum, Leech and Svartvik1985: 485), predominantly in direct speech.

1. (4)

   Ic eom soðlice romanisc. and ic on hæftnyd hider gelæd wæs.

   ‘I am truly a Roman, and I was brought hither in captivity.’

   (LS 8 (Eust) 344, DOEC)

In Old English narrative prose (i.e. not in direct speech), soþlice mostly serves a different function, that of a sentence adverbial, in which it loses much of its original meaning, extends its scope from the phrase level to at least the sentence or even discourse level, and at the same time develops a metatextual function as a discourse marker, e.g. indicating the beginning of new episodes in the gospels or highlighting certain sentences on the local level of discourse (for a fuller account and examples, see Lenker Reference Lenker, Fischer, Rosenbach and Stein2000, Reference Lenker, Fischer, Rosenbach and Stein2007).

5.3 ME forsooth

Soothly (< OE soþlice) continues to be highly frequent in Middle English in all of its Old English functions, i.e. manner adverb, emphasizer, and discourse marker (see Lenker Reference Lenker, Fischer, Rosenbach and Stein2007: 83–9). From the beginning of the thirteenth century onwards, however, forsooth begins to be frequently attested in similar contexts and functions in all dialects of Middle English (see OED and MED s.v.; Appenzeller-Gassmann Reference Appenzeller-Gassmann1961: 47–8; Swan 1988: 259–62). It appears in various regional and orthographical forms, such as (to list only the main variants) for soþ and for soþe and also the univerbated forms forsoþ, forsoþe, forsuth, for-sute, and Southern vor zoþe, vorzoþe (OED s.v. forsooth, MED s.v. forsoth). It is used in all of the functions attested for OE soþlice, but is primarily employed in an emphasizer function in direct speech – often in answers to questions, both positive and negative ones (see also MED s.v. forsoth 2.b).

1. (5)

   Boece. ‘Ye/Yis, forsothe’, quod I

   (Chaucer, Boece3, pr11.203; pr12.57, CME)

2. (6)

   ‘Nay/No forsothe’, quod I

   (Chaucer, Boece3, pr12.150; 4, pr2.100 etc., CME)

In this function, it is not only attested in translation texts but is also quite frequent in original Middle English prose, for example in Malory's Morte D'Arthur:

1. (7)

   ‘Ye forsothe’, said the queen . . .

   (Book 13, cap. vii; etc., CME)

This predominant use of forsooth in interactive dialogue as an emphasizer is also reflected in its further history, when – as noted in Samuel Johnson's Dictionary – it even becomes an honorific, a ‘word of honour in address’ (Johnson Reference Johnson1755: s.v. forsoo'th; for these uses, see Lenker Reference Lenker, Kornexl and Lenker2003: 262–4, 283–6).

5.4 Etymology and morphological make-up