Exploring Creativity and Extravagance: The Case of Double Suffixation in English

Abstract

There has been a recent focus in studies of English morphology on the concept of extravagance as applied to word formation, and on the interplay between extravagance and creativity. This article examines this issue, taking as a test case the phenomenon of double suffixation of phrasal verbs. While double-ER suffixation (fixer-upper, helper-outer) has attracted substantial interest in the literature, less has been said about other suffixes. This article provides data that demonstrate that double suffixation occurs with -ERY, -AGE, -EE and -ABLE suffixes. As such, double suffixation can be seen as a genuine word formation template, rather than a phenomenon restricted to a single suffix. Furthermore, examination of the data enables a reflection on the interactions between the concepts of productivity, creativity and extravagance. We propose to see these concepts not just as three overlapping notions, but rather as points on a creativity scale. To underscore the parallels between them, we propose to see them successively as F-creativity, E-creativity and X-creativity.

1. Introduction

Contemporary English allows for what have been named fixer-upper nominalisations, among other terms, which have been the focus of some attention in recent years (Chapman 2008; Walker 2009; Lensch 2018, 2022—see the following section for further references). This can be seen as a word formation pattern affecting phrasal verbs. This pattern appears to involve redundant suffixation of the nominalising -ER suffix on both the verb and the accompanying particle. These forms raise a number of interesting questions, in particular with regard to what we can learn from their morphology. There are two aspects, however, which remain relatively little explored, and it is to these to which this article turns. The first is indeed morphological in scope: while the redundant -ER has been well discussed, there is little research into whether this is a pattern available for other suffixes1. The second builds on this to make a more general theoretical contribution to the notions of creativity, productivity and extravagance, and the interplay between them. In so doing, we wish to see not so much an ‘overlap’ (Eitelmann and Haumann 2022, p. 4) between them, but a spectrum which allows us to see this word formation process in dynamic terms.

This article will touch on each of these questions, to make the following claims: first, that double suffixation is redundant on a surface level, but the status of the two suffixes is not identical. While both appear to be the result of typical morphological processes, only one is meaningful. The other is a semantically vacuous marker of ‘linguistic playfulness’, a pre-theoretical term we use here to avoid dealing with the extravagance–creativity–productivity cline we posit in the second claim. The second claim, therefore, is that what we refer to as double suffixation in English is a word formation process which illustrates a process of stabilisation running from extravagance, through creativity, to productivity. In other words, double suffixation used to be extravagant (and still is, for many cases) and now (at least for double-ER) occupies a different area on the creativity–productivity scale. It is our contention that a certain degree of haziness in the literature between the terms productive, creative and extravagant can be if not resolved, at least satisfactorily illustrated by using double suffixation as an example. Furthermore, we see this as tying in with Eitelmann and Haumann’s (2022, p. 5) idea of extravagance which is “essentially conceptualised as a trigger of language variation and change”. This conceptualisation requires a dynamic representation perhaps absent from the metaphor of the ‘overlap’.

The article is structured as follows: Section 2 will set out a brief review of the existing literature on the double -ER pattern. Section 3 will add data to the discussion, by demonstrating the existence of similar forms with various suffixes. In Section 4, we discuss the notions of productivity, creativity and extravagance as they pertain to these examples and we propose a dynamic ‘cline of creativity’. The Section 5 concludes.

2. Previous Treatments of Double-ER Suffixation

The process by which phrasal verbs2 in English can be nominalised using what superficially looks like a reduplicated -ER, as in picker-upper or helper-outer, has been the focus of some attention in recent years and constitutes a “novel morphological template” (Lensch 2022, p. 76), which can be represented as follows:

(a)

VERB + -ER + PARTICLE + -ER

pick + -ER + up + -ER = picker upper

In this pattern, each -ER seems on the face of it to be the conventional nominalising deverbal agentive suffix -ER found in countless derivatives in English, such as teacher, baker or helper. This “novel template” stands alongside others for the derivation of nouns from phrasal verbs, such as

(b)

PARTICLE + VERB + -ER

by + stand + -ER = bystander

(c)

VERB + -ER + PARTICLE

pass + -ER + by = passer-by

(d)

VERB + PARTICLE + -ER

come + out + -ER = come-outer ‘a person who withdraws from an organization, institution, etc., as a result of religious or political principles’3

These forms appear to have been around for at least a century or so. The first Oxford English Dictionary attestation of picker-upper dates to 1913, and other examples to be found in the Oxford English Dictionary (OED), attesting to a degree of stabilisation (lexicalisation), are maker-upper (1936), opener-upper (1941) and tearer-downer (1942)4. They have also been the object of occasional comments in the literature (Walker 2009 and Lensch 2022 provide good coverage). It is only recently, however, that the phenomenon has attracted more detailed scholarly attention. Chapman (2008), referring to “picker-upper constructions”, seems to be the first article-length discussion of the topic, and here they are analysed as the result of ‘double marking’ of an -ER suffix, perhaps the mid-point of a shift in progress from a historically preferred picker-up pattern towards a newer pick-upper model. Walker (2009) also refers to ‘double suffixation’, but seeks to go beyond a purely morphological reading of the pattern, emphasising the importance of the metrical pattern of four-syllable structures with alternating strong and weak suffixes and speculating that euphony is a factor in the development of the pattern. He draws parallels with examples of ablaut reduplicates very often found in children’s rhymes (Arleo 2001), and offers a reflection therefore on the fact that the two suffixes are not performing the same task. One is a conventional deverbal nominaliser, while the other is important auditorily or visually, but serves no semantic or morphological purpose.

This characterisation of the -ER suffix as deverbal is an oversimplification, of course. As is well known, -ER is also a nominaliser of adjectives (lone + ER = loner) and a deriver of nouns from nouns and numerals (Iceland + ER = Icelander; 4 tonnes + ER = 4-tonner). Crucially for what follows, and something that the literature on the subject does not seem to have picked up on, is that -ER is also a historically ‘jocular suffix’. The OED describes it as “[u]sed to make jocular formations on nouns, by clipping or curtailing them and adding -er to the remaining part, which is sometimes itself distorted”. Among the earliest instances are footer (from football), rugger, soccer, fresher, bedder, bonner, brekker or ekker. An example of a proper name is Bodder (‘Bodleian’).

It can hardly be a coincidence that the -ER suffix used apparently redundantly to give trochaic forms such as breaker-upper and shutter-outer, is the same used to form ‘jocular’ nouns which are also overwhelmingly trochaic in metre, and exclusively so in the examples given in the OED. This also anticipates the findings, set out below, for the closely related -ERY suffix. This serves to reinforce the argument made in Walker (2009) on the influence of euphony and the fact that, to come back to fixer-upper, we are not dealing with reduplicates. This goes counter to the analysis set out in Cappelle’s (2010, p. 335) treatment of the pattern, which refers to “reduplicate forms”. Cappelle is also dismissive of an analysis based on euphony or ‘word-internal rhyme’, seeing it as ‘linguistically naïve’ (Cappelle 2010, p. 348). However, Cappelle is clear that the suffixes do not have the same status and later reflects on the metrical motivation for the pattern, agreeing with Miller (1993, pp. 132–33) that the “first -er is essentially vacuous”.

Lensch devotes two papers to the issue (Lensch 2018, 2022), and is the first author to link the phenomenon to the feature that has come to be known as “extravagant morphology”. She refers (Lensch 2022, p. 76) to this ‘multiple attachment’ as a ‘novel morphological template’ and carefully sets out a number of the interesting morphological puzzles posed by these V + -ER + particle + -ER forms. They attest to the fact, for instance, that reduplication is perhaps more pervasive in English than has generally been held to be the case. These forms violate Williams’s (1981) ‘Right Hand Head Rule’ (the head of a lexeme is the right-hand element), challenge the ‘monosuffix constraint’ put forward by Aronoff and Fuhrhop (2002, words in English tend not to have more than one Germanic suffix) and contravene Rohdenburg’s (2003) ‘Horror Aequi Principle’, a purported universal tendency to avoid adjacency of identical elements. If, however, we accept that we are not, strictly speaking, dealing with two suffixes but with one suffix and either an ‘echo’ of the same suffix, or an anticipation thereof, the first two of these apparent contradictions are resolved. The question as to whether the ‘empty’ suffix is an echo or an anticipation will not detain us greatly here. The question, which is ultimately related to the wordhood and/or headedness of phrasal verbs, is of course important. Inasmuch as a form such as washer-upper seems to be admitting of double suffixation, both on the verbal head and on the particle, one conclusion is that the verb wash up has two heads, or that there is at the very least uncertainty as to where the head is. The competing forms, washer-up (OED entry) and wash-upper5 resolve the headedness issue in conflicting ways. Suffice it to say that since the VERB + PARTICLE + -ER (pick-upper, ask-outer) form is the commonest, a pattern we see repeated for the other suffixes examined below, we might wish to claim that the ‘real’ derivative morpheme is appended to the particle, and the first stands in anticipation thereof.

However, in what follows, what is at stake is our contention that there is indeed a principled difference between the two suffixes at play, irrespective of the debate between an echo effect or an anticipatory effect. In so doing, we follow Newell (2009, p. 203), who says “[t]he morpheme is only present once in the narrow syntax and it is a purely phonological and interpretive effect that causes it to be overtly realized twice”. In other words, one of the -ER suffixes is a ‘genuine’ nominalising suffix. The other is not a reduplicate of the first, rather it is a ‘semantically vacuous’ or ‘pseudo-morphological’ (Walker 2009) marker of jocularity, as the OED has it, or of colloquialness, to follow Bauer (1983, p. 289), who says “they [double -ER derivatives] tend to feel very clumsy, and as a result tend to be used mainly in colloquial speech”. For our theoretical purposes, the double -ER suffix is a “marker of extravagance”.

It might be argued (as one anonymous reviewer, to whom we are grateful, pointed out), that it is not possible to falsify, in the Popperian sense, the claim that one of the suffixes is vacuous or semantically non-contributing, and that it therefore serves the purposes of extravagance. Such a falsification attempt would require the experimental demonstration that there is no semantic difference between single- and double-affixed forms, such as pick-upper vs. picker-upper. We run the risk, this argument runs, of overstating the case for double affixation being a marker of extravagance. Recall, however, that we make the statement not as an axiom, but as a hypothesis to be explored. Furthermore, as the discussion below will make clear, we will want to claim that extravagance is a fundamentally dynamic notion, and that such vacuousness is typically to be found at the early stages of a language shift and not necessarily as a permanent feature thereof.

3. Further Data

The vast majority of work on double phrasal verb suffixation thus far has focused on the -ER suffix, as mentioned in the opening remarks. This focus, while it raises a number of interesting questions, cannot lay any claim to be the harbinger of a genuine process of word formation in English, absent similar examples being adduced for other suffixes. This section thus provides a data set with that end in mind.

Searches were conducted on the Internet for examples of words derived from phrasal verbs by means of double suffixation. No claim of exhaustivity is made for the searches conducted, which are intended to be illustrative and exploratory, rather than definitive. The search protocol was as follows: a Google search6 was conducted using, for example, the wild card string “the * uppery”, in order to produce results where the word the was followed by one word, itself followed by uppery (and similarly for other particles, such as out, down etc, and likewise similar for other suffixes, such as -EE and -AGE. The full list is provided below. Successive searches were conducted for “the * outee”, “the * downage” etc. These results were then manually checked for phrasal verbs, as the majority of hits were noise—for instance, Uppery is the name of a chain of nightclubs, outery was a frequent OCR mistake for outcry in digitised texts, uppage is a local name for the Garcinia gummi-gutta shrub, etc. Once a potential phrasal verb combination was spotted, a further search was conducted for each of the patterns we have already seen for -ER derivatives, namely:

(a)

V + particle + suffix

pop + out + -ER = pop outer

(b)

V + suffix + particle

pop + -ER + out = popper out

(c)

V + suffix + particle + suffix

pop + -ER + out + -ER = popper outer

This third pattern lies at the heart of our further discussion, while patterns (a) and (b) were sought out for comparative purposes and to constitute a corpus for possible further research.

The first suffix we examine is the -ERY suffix, which shares first and foremost a nominalisation function with -ER. It might be objected that -ERY is more properly analysed as the -Y suffix appended to a V + ER nominalisation, as in bakery or ironmongery (see OED 2023, s.v. “-ery, suffix”, July 2023), rather than an independent suffix appended to a verb. We hold first that the ‘conventional’ nature of a morpheme is of little import, due precisely to the redundancy we are investigating. Further, -ERY is similar to -ER in its polyvalency, referring to a place (fishery), goods (pottery), collective senses (machinery), negative characteristic (popery) and place for housing animals (piggery). Third, and intriguingly, -ERY shares with -ER a 19th-century vogue for nonce formations, such as Colinderies (Colonial and Indian Exhibition) and Dukeries (tract of country occupied by the great ducal estates in Nottinghamshire and North Derbyshire). For all these reasons, the parallels with -ER are interesting to examine. The results are set out below.

The phrasal verbs found in combination with the -ERY suffix were fuck up, suck up, shut up, make up, pop up, cock up, scuff up, freak out, hang out, make out. All of these verbs were found exhibiting the pattern V + particle + suffix, as in suck-uppery or make-outery:

(1)

Besides the expected obsequious suck-uppery here (and this guy has a really strong suction game), I was struck by the need to add “robust health” to this statement about Murdoch7.

(2)

A bit of hot fluff rated T for make-outery and various states of undress8.

Only two of the verbs were found with double -ERY suffixation, fuck up and suck up:

(3)

GOD doesn’t forgive this level of fuckery-uppery… you’ll see…9

(4)

Interesting to see @KTHopkins switch her suckery uppery so firmly towards Russia though10

Only one exhibited the V + suffix + particle-pattern, namely fuck up:

(5)

I send this in, CentComm’ll want to do a gigantic fuckery-up all over the DMZ, and that’ll muddy the waters11.

Table 1. Phrasal verbs and -ERY nominalisation.

	V + Part. + ERY	V + ERY + Part. + ERY	V + ERY + Part.
fuck up	3980	35	1
suck up	1700	1	0
shut up	169012	0	0
make up	39	0	0
pop up	894	0	0
cock up	60	0	0
scuff up	1	0	0
freak out	310	0	0
hang out	111	0	0
make out	149	0	0

presents the data as follows.

Moreover, we examined two further nominalising suffixes, namely -AGE and -EE. The further searches are only exemplified for the double suffix pattern, because only these are germane to the discussion below, as show in (6) and (7) on the basis of the examples poppage-uppage and breakee-uppee, respectively. We provide the figures for the other forms in the following tables (Table 2 and Table 3).

(6)

Seed poppage uppage! Here they come CHOO CHOO!!! So far 4 of the 6 have completely broken the surface13.

(7)

It lays all the responsibility on the breaker-upper as opposed to breakee-uppee14.

Table 2. Phrasal verbs and -AGE nominalization.

	V + Part. + AGE	V + AGE + Part. + AGE	V + AGE + Part.
fuck up	224	3	6
shut up	4	0	0
pop up	6	2	0
hop up	2	0	0
freak out	attested15	3	3
make out	attested	0	0
pop out	attested	2	5
knock out	attested	0	0
hang out	attested	4	0

Table 2. Phrasal verbs and -AGE nominalization.

	V + Part. + AGE	V + AGE + Part. + AGE	V + AGE + Part.
fuck up	224	3	6
shut up	4	0	0
pop up	6	2	0
hop up	2	0	0
freak out	attested15	3	3
make out	attested	0	0
pop out	attested	2	5
knock out	attested	0	0
hang out	attested	4	0

Table 3. Phrasal verbs and -EE nominalisation.

	V + Part. + EE	V + EE + Part. + EE	V + AGE + EE.
fuck up	4	0	0
break up16	681	1	0

Table 3. Phrasal verbs and -EE nominalisation.

	V + Part. + EE	V + EE + Part. + EE	V + AGE + EE.
fuck up	4	0	0
break up16	681	1	0

Finally, we also examined the suffix -ABLE. The pattern V + particle + -ABLE is quite well attested. The OED provides entries for put-up-able-with, knock-upable and lock-upable, and a quick Google search provides examples, with common collocates, such as fuck-uppable, curl-uppable (chair), un-mess-uppable (hair), stay-uppable (TV show), top-uppable (Oyster card), write-uppable (offence), snap-uppable (ticket), zip-uppable (hoodie) and many more besides.

The double suffixation pattern, while much less common, as we saw for the -ER nominalisations, is also attested. It seems, however, to have featured a great deal less in the literature. An exception to that is Los et al. (2012, p. 134) who, talking about suffixes on ‘verb particle complexes’, give the example of lookable-uppable. The question appears as a footnote in Farrell (2005, p. 9) and as an example which is not discussed in Newell (2009, p. 198). Some examples are given in (8) and (9). The whole corpus can be summarized in

Table 4. Phrasal verbs and -ABLE adjectivization.

	V + Part. + ABLE	V + ABLE + Part. + ABLE	V + ABLE + Part.
look up	816	60	46
fuck up	631	2	0
break up	213	2	0
pick up	8835	238	232
top up	929	8	5
back up	320	5	8
use up	49	1	0

below.

(8)

And once you know the author, the character whom the author used as mouthpiece is lookable-uppable, for them as are interested17.

(9)

Maybe there’s a way to have the email stored on my hosting provider (as a backable-uppable database, not using webmail)18.

The examples with various suffixes show that this phenomenon is not restricted to -ER suffixation. Examples with double -ERY, double -AGE and double -EE suffixes prove that we are not only talking of a morphologically limited phenomenon, rather a word formation process. Moreover, the numerous attestations of double -ABLE show that the process is not restricted to nominalisation; rather, it is a frequent word formation pattern that involves the double appearance of a suffix attached to both members of a phrasal verb. More on this process will be discussed in the following section.

4. The Notions of Creativity, Productivity and Extravagance

The notions of productivity and creativity, though not without their polysemy, are relatively well-established terms within morphological theory. They go back at least to Chomsky on language in general, and more specifically to generative morphology (see for example van Marle 1984). Interestingly for the thesis we set out below, creativity is a notion that has lagged behind that of productivity in the literature. Many a book tackles the thorny issue of just what morphological productivity should be understood to be (Baayen 1989; Plag 1999; Bauer 2001 being just three examples), whereas references to ‘morphological creativity’, ‘creative coinages’, ‘lexical creativity’ and the like tend more often to form the focus of a chapter or of shorter sections. More recently, creativity with emphasis on word formation and word interpretation has been investigated by Körtvélyessy et al. (2022). Additionally, we have seen the emergence of the concept of morphological extravagance, with the work edited by Eitelmann and Haumann (2022). This is a very new addition to the morphological pantheon, and naturally therefore much remains to be worked though. The present paper seeks to contribute to the discussion of these notions, and the links that bind them to the concept of productivity. In the following, we successively discuss the interplay between productivity, creativity and extravagance, before coming to illustrate this discussion with the double suffixation data, and in particular certain telling examples.

4.1. The Productivity–Creativity Overlap

We take productivity to be an uncontroversial term and define it as the unnoteworthy replicability of a given pattern or schema for word formation. This comes very close to a Chomskyan view of creativity, which Chomsky himself recognised as “idiosyncratic” (Chomsky and Foucault [1974] 2006, p. 19). It makes creativity ‘normal’ and strips away from it any noteworthiness which we might instinctively associate with the word. Chomskyan creativity is a universal feature of all humans and is the consequence of the regular implementation of productive rules to generate new sentences. It can be seen, therefore, as overlapping, perhaps being indistinguishable from, a notion of productivity.

Another view of the relationship between creativity and productivity takes its roots in a criticism of this Chomskyan ‘normalisation’ of creativity. This view sees creativity in contrast to productivity rather than the two being in superposition. Productivity is the bedrock to creativity. By that, we mean that there can be no creativity without productivity, the ordinary, quotidian means by which a speaker can create unnoteworthy utterances which are novel only in the trivial sense that they have never been uttered before. We see creativity as the capacity to generate linguistic material which is, in some sense, noteworthy. At the morphological level, we see creativity as pertaining to words that “draw attention to themselves. They will be perceived by the native speaker as odd, amusing, repulsive, or otherwise remarkable” (Lieber 1992, p. 3).

This is not an unproblematic view. A word can quite clearly be both repulsive and entirely uncreative—repetitive vulgar language would surely fall into this category—but this approach has the virtue of highlighting the importance to the distinction of ‘noteworthiness’. This, too, is not without its problems. Is noteworthiness to be ascribed to the speaker, as an intention? It is never safe ground for a linguist to be travelling, if his or her theoretical construct is based on unknowables such as intention. Or is noteworthiness an effect produced on the receiver? A combination of the two?

This tension is captured nicely by Sampson (2016) in the distinction he makes between F-creativity (for ‘fixed’) and E-creativity (for ‘enlarging or extending’). The former matches the Chomskyan perspective of the generation of new content from a ‘fixed’ range of possibilities, whereas the latter covers creativity as “enlarge[ing] our understanding of the range of possible products of the activity” (Sampson 2016, p. 19). We see these two notions as two points on a scale of creativity as shown in Figure 1 below.

The advantages of representing matters on a scale such as this, rather than as two opposing blocks, are not merely decorative. It underscores the fact that our vision of the two categories is fundamentally dynamic and transitional, across a number of areas. First, it captures the intuition that a formation can be F-creative in one context and E-creative in another—synchronic transition. Second, it captures the fact that a formation can be F-creative at one point in the history of the language and E-creative in another—diachronic transition. It helps represent one of the explanatory issues facing us from a language change perspective. How does something presumably fleeting and individual (the E-creativity end of the scale) come to be ‘normalised’ as part of the collective knowledge of the speech community, and thus move to the F-creativity end of the scale? Last, it captures the intuition that what is F-creative for one speaker-listener dyad might be E-creative for another—context dependency.

Furthermore, this creativity scale can be bidirectional. When a speaker is overtly creative and playful, then they occupy a space along the scale which is different from when they are less so. When an initially extravagant structure (cf. Figure 2 below) becomes more E-creative and then potentially more F-creative, then it moves back along the scale. It also presents the advantage of being extendible, as we demonstrate in the next section, and, by eschewing the notion of productivity in favour of F-creativity, we emphasise that we are dealing with two points on a scale and not two contrasting notions.

Seeing the distinction between F-creativity (productivity) and E-creativity as a cline is by no means innovative. Authors such as Hohenhaus (2007, p. 15), “the distinction between productivity and creativity—a cline”, Konieczna (2012, p. 12) “it seems justified to perceive the difference between creativity and productivity in terms of a cline and prototypical values rather than definite categories” and Beliaeva (2019: introduction) “this study aims to […] develop criteria that could be applied to corpus data to characterize particular examples of blending as points on a cline from word creativity to productive word formation”, among many others, make the same point. We believe that what sets our proposal apart from others is a. its multi-factorial nature (synchronic and diachronic, style and register dependency, nonce and conventionalised uses), b. its bi-directionality and c. its extension to cover a new addition to the pantheon of morphological theory, the issue of extravagance, to which we now turn.

4.2. The Creativity–Extravagance Overlap

The introduction of the notion of extravagance into the picture can only make sense if there is dissatisfaction with the F-creativity and E-creativity cline. In other words, if there are pertinent linguistic phenomena which indubitably exist, but which have characteristics which render the term E-creativity insufficient. Reading some authors, we might be forgiven for questioning this need. Bergs (2019, p. 180), for example, holds that the only form of linguistic creativity is that of ‘aberration’, i.e., when language production is absolutely unpredictable and does not follow any obvious rules or constraints, which he says applies to extremely rare cases. In decoupling creativity almost entirely from rule governance, Berg espouses what is perhaps rather an outlying view. Munat (2007) sees the distinction between (lexical) productivity and creativity as hinging on the latter’s use as an attention-seeking device, a term which has proved extremely fertile for further research.

However, we shall follow a more recent trend in positing that there is a need to see the creativity–productivity dichotomy as insufficient, and that we need to see the cline extended. We will take seriously the notion that there is a need to talk about extravagance, as defined by Eitelmann and Haumann (2022, pp. 2, 5), and that this notion is a ‘trigger for language variation and change’. We shall attempt to reconcile the notions of extravagance and creativity, seen to overlap by Eitelmann and Haumann (2022, p. 4), without the nature of the overlap being extensively explored. In so doing, we also keep in mind Körtvélyessy et al.’s (2022) view, according to which creativity in word formation is seen as a manifestation of the human’s cognitive creative potential and so it is a feature of a language community.

The use of the term extravagance goes back to the fifth of Haspelmath’s (1999, p. 1055) maxims, namely ‘talk in such a way as you will be noticed’. This has been picked up by many scholars, and developed into the notion of linguistic forms as “attention-seeking devices” (Hohenhaus 2007, p. 23). However, once again, we see an overlap, not to say a contradiction, between those authors, also such as Lalić-Krstin (2018) or Ahmad (2018), heralding attention-seeking devices as a sign of creativity, using the terms creativity and attention-seeking more or less interchangeably, and those, like the present authors, entertaining the need for a distinction.

By already introducing a cline, we have seen that creativity, following Kecskes (2016, p. 3) is a “grade phenomenon ranging from the more conventional and predictable to the less conventional and predictable”. We propose to extend the scale a little further, and introduce the idea of ‘stretched creativity’ on a dynamic cline. This enables us to go beyond the idea of the frontier between creativity and extravagance being fuzzy or difficult to define, by eliminating the notion of a frontier altogether. There is no frontier, and therefore no overlap; rather, both are different points on a scale of creativity. We propose here the term ‘X-creativity’.

There are several reasons for introducing this term. X-creativity can be seen to refer to the notion EXtravagance, which it in some sense encompasses. One other reason for preferring the term X-creativity is that it perhaps also encapsulates the initially fundamentally individual, and often unknowable origin, of an extravagant, or X-creative form. One of the issues facing the proponents of ‘extravagance’ is the theorisation of how one goes from what is fundamentally an individual notion to something that is collective. The term ‘extravagance’ itself seems to underscore the individuality, the marginality of the phenomena under study. If it remains an individual phenomenon, it is of very little theoretical value—an aporia. For instance, the fact that we might occasionally find extraordinary one-offs such as foul weather driver maker funners of, seen in (10), tells us very little. However, if we begin to witness a shift from extravagance to a more collective and process-based phenomenon, then things become a little more interesting.

(10)

“Perhaps some of these foul weather driver maker funners of should move back to the Midwest where people enjoy ice skating with their cars” (Yamabe n.d., p. 51)

The reason for wishing to continue to use the word ‘creativity’ is neatly captured in a definition of extravagance provided by Eitelmann and Haumann (2022, p. 3), as “a term capturing any norm-deviating and rule-bending language use that may serve as the entrance door to variation and change”. The important issue at stake here is that there is explicitly a link between extravagance and language change. This brings a diachronic, or at the very least a transitional perspective, into the picture. The distinction between diachronic and transitional is key here, as the argument might be made that oftentimes with the phenomena labelled as extravagant, we are perhaps not dealing with rules changing as a result of ‘rule-bending’; rather, we are observing a greater social acceptance of rule-bending. In other words, not language changing (the ‘rules change’), but social change (the rules become more ‘bendable’). It is this transitional, as opposed to purely diachronic, perspective that the visual representations we propose seek to capture. Before going back to the double suffixes to illustrate this cline, we introduce one further point on the scale.

4.3. The Extravagance–Error Overlap

Another potentially problematic feature of the definition of extravagance provided by Eitelmann and Haumann (2022, p. 11) lies in the relationship with the notion of error. They say “Extravagance is understood as the stretching of processes beyond their usual boundaries, often in tandem with violations of well-established rules and constraints, without generating ill-formed output”. The issue here is that of ill-formedness, because it is used as if that were an objective and unproblematic notion. Yet, as we know, ill-formedness is in the eye of the beholder. If it were not, the above definition would be internally contradictory, because it is hard to see how the violation of a constraint could not, by definition, produce an ill-formed notion. Put otherwise, it is important when talking of ‘rule-bending’ to see that it is not just a matter of to what extent rule X is ‘being bent’, but crucially, to what extent the speaker believes s/he to be bending a rule and the extent to which the receiver accepts the bending of the rule.

The question of the border between “when a novel utterance is seen as creative and when it is considered as wrong” (Hoffmann 2020, p. 1) is much vexed and we shall not rehearse it much here. We shall content ourselves with dispensing with the notion of ‘border’ (the ‘thin line’ of the title of the paper to which Hoffmann is responding) to place it on the same scale we have been using up until now, as Figure 3 shows.

This seems to capture intuitions such as what is ill-formed for one person, at one time, in one dyad, in one context, in one register, may be X-creative for another person, at another time, etc. It will by no means always be easy to place different speech events on the scale, but we will point to a number of metalinguistic indications in the data, specifically hesitations and self-corrections, as well as the imperfect bending of rules, to tease this out.

4.4. Discussion

How are we to go about using the data provided in Section 2 to illustrate the multi-factorial cline? In other words, how are we to determine, in a manner which is not entirely an ex-cathedra pronouncement, to what degree a given form is X-creative, or F-creative, for instance? We seek to show that certain clues can help us achieve this aim, among which are metalinguistic comments, hesitations and self-corrections. Instances of our corpus are retrieved for this purpose.

While frequency is not the be-all-and-end-all, there is little doubt that there are frequency effects at play in what we see as a progressive entrenchment of the double -ER pattern. Much of the output is lexicalised, and there are many thousands of examples to be found across a whole variety of genres, from Internet forums to broadsheet newspapers, from Reddit to any number of novels. One might therefore be forgiven for claiming that double -ER has become fully entrenched, and therefore F-creative, as the single -ER indubitably has. However, we can point to occasional hesitations as signs that the process is not complete. It must be admitted that the internet data we use here is potentially problematic, inasmuch as there is no cast-iron proof that the persons producing the form are native speakers, and hesitation may be a feature in particular of non-native speech. However, in this case we are dealing with extravagant linguistic forms that are almost certainly the preserve of native speakers, who also show (by hesitating) conscious use of the language and uncertainty as for the ‘correct’ word formation rules. Examples include

(11)

Guys are usually the asker outer (is that a word?). (Walker 2009, p. 11)

It is inconceivable that a speaker would hesitate as to the “wordness” of a fully F-creative form, and so therefore we would wish to place this form somewhere between the E-creative and F-creative nodes on the scale.

Similar evidence can be found in orthographical clues which point to a lack of stability in the morphological status of the -ER, such as:

(12)

Are you a good adder-upper-rah. (Walker 2009, p. 11)

It must be owned that here, we do not have a double -ER, but in fact a triple -ER. While this therefore, in all rigour, cannot be put forward as a demonstration of the positioning on the scale of the double -ER, it does provide further evidence of the ‘semantic vacuity’ of the ‘extra’ suffixes. The third ‘suffix’ is not analysed by this speaker as being a repetition of the first or second, so a different spelling is licensed. Such triple suffixations would surely find themselves around the X-creative node, on this evidence.

Similar signs of uncertainty are to be found with other suffixes contributing to the same pattern. This suggests to us not so much an attempt to be extravagant, since this latter is purported to involve rule-bending, surely a deliberate process—can one be unwittingly extravagant? Here are some instances of hesitation involving phrasal verb suffixation. We find one such example (freak out-age/freakage out?) in the following discussion on a Wikipedia help page.

(13)

Question: Would anybody freak out, for example, if i Wikilinked the 24,000 instances of “blame” throughout Wikipedia?

Answer: Freak out-age (freakage out?) would be very likely if you did that, because of that guideline on the limits of overlinking in individual articles19.

The speaker here opts for one form, but offers up a second as a potential alternative, without comment, and this can be analysed in two contradictory ways, both of which point to the need, once again, for a cline-based analysis. First, the hesitation could be seen as constituting almost the opposite of an attention-seeking device, a genuine uncertainty about the appropriate word-formation process to use and an insecurity about making a potential mistake and appearing ‘too extravagant’. Alternatively, it could be a sign of what we might want to term ‘paradoxical hesitation’, a drawing of attention to the fact that the speaker masters two possible word-formation processes and is aware of the potential of using both, successively, as an attention-seeking device. Here, it is not so much the competing forms in themselves, but the confident use of two patterns occupying slightly different places on the cline, which constitute the attention-seeking device. As a result, we would wish to place this somewhere to the right of the E-creative node.

A similar example, this time with a double suffix, is the following:

(14)

Are you the Breaker-upper or Breaker-ee? (Breaker-upee?)20.

The process at play here in (14) is identical to that in (13), the offering up of an initial output, which is immediately questioned with a second, and an explicit recognition of the speaker’s uncertainty. What is more interesting here is that in neither case does the output seem to be well-formed. The first proposal appends the suffix -EE directly to the form breaker, written with a hyphen as a signal of the speaker’s morphological awareness. This is entirely anomalous, if subjected to a conventional morphological analysis. If seen, however, not as a performance error, but as a sign of a speaker’s uncertainty of the word formation process, which is a definite sign that we are a long way from F-creativity, it is more interesting. The proposed correction, breaker-upee, does not resolve the issue, and it is this kind of ‘unsuccessful correction’ which enables us to posit a position for this pattern on the creativity scale as being somewhere between X-creativity and ill-formedness.

Examples of explicit comments with other double suffixes also exist, such as:

(15)

Of all the people who said it’s true in your life, how many of them have provided any kind of lookable-uppable source? Exactly. And lookable-uppable is a real word21.

What we take to be happening here is a form of anticipation that a receiver might deem the double -ABLE to be ill-formed, or at the very least inappropriately extravagant. Our speaker seeks to dispel this, perhaps rather defiantly, and places the form, arguably, between the X-creative and E-creative nodes on the scale.

5. Conclusions

In this paper, we have dealt with several challenges, and we have reached a number of conclusions. Providing novel and naturally occurring data, we claimed that the discussion should not be just around double -ER suffixation. Examples with double -ERY, double -AGE, double -EE and double -ABLE suffixes prove that we are not only talking of a morphologically limited nominalisation phenomenon, rather a word formation process. Moreover, we have moved away from the analysis that we are dealing with cases of reduplication. While one of the suffixes is a nominalising or adjectivising suffix, the other is semantically empty, a marker of extravagance. Our contention is that double suffixes are creative to varying degrees, in a paradigmatic fashion across phrasal verbs as a whole. However, there seems to be a real correspondence between the familiarity/vulgarity of the phrasal verb (e.g., fuck up) and the double suffixation, which makes us think that there is a link between creativity and familiarity.

Our aim was also to go beyond extravagance as an observation of it being a trigger of language change and variation and place the notion onto a creativity scale (scale 2). Our data led us to establish a scale extending from ‘Ill-Formedness’ to ‘F-creativity’ from fully productive forms to almost ill-formed structures (scale 3). We introduced this scale in order to resolve the overlap issues and the haziness between the notions of creativity, productivity and the newcomer in the theory of word formation, the notion of extravagance. We see extravagance as a ‘stretched creativity’. We claim that extravagance (X-creativity, in our terms) begins as individual, collectivises to become creativity (E-creativity) and stabilises to become productivity (F-creativity).

The proposed creativity scale is dynamic and transitional, synchronically as well as diachronically. It can include structures that are context-dependent, in the sense that what is F-creative for one speaker–listener dyad might be E-creative for another. Moreover, it is bidirectional since structures can move towards to the one end while others can move towards the other end. X-creative structures can move to the ill-formedness end, as hesitation instances had shown, or towards the F-creativity end. The more productive, the less ‘noteworthy’ a structure becomes. We have traced both directions of that process using the examples of double suffixation in English. Phenomena move along this scale, according to criteria such as frequency, context of use, ludicity, predictability and comprehensibility. Language change plays a crucial role in this process.