‘There’s far too much arty-farty, namby-pamby, hoity-toity, wishy-washy, lardy-dardy, sun-dried tomato eating, decaffeinated fannying about!': Reduplicative compounds and other iterative sequences in English

ABSTRACT

The aim of the paper is to (re-)evaluate the treatment of reduplicative expressions as compound words in nearly all previous work on English. Examples of reduplicative expressions are listed, out of context, and they are labelled as compound words, without providing proper motivations as to why it makes sense to treat them as compounds, in the sense of lexical items that are formed of other lexical items. The paper provides plenty of examples of how the reduplicatives listed in previous work are actually used in present-day English and examines the functions and interpretations of such expressions. The paper shows that most reduplicative expressions are not compound words. Some of them are words in their own right that cannot be divided into meaningful constituents. Some may have a more compound-like reading, alongside a clearly non-compound-like reading; for most such items, the latter function and interpretation seem to be most common. A large majority of alleged reduplicative compounds display clear non-compound-like behaviour. It will be shown that these expressions are likely to be sequences of words that are created by the same (syntactic) process/es that iterate or stack ‘ordinary’ nouns, adjectives and verbs in English.

KEYWORDS:

• Reduplication
• repetition
• iteration
• compounding
• English

1. Introduction

Repetition or reduplication of linguistic material occurs commonly in natural language and communication. According to Jespersen (1942: 173), it is ‘natural to all human beings and is found very often in all languages as a means of strengthening an utterance’. Sapir (1921: 76) has argued that ‘nothing is more natural than the prevalence of reduplication’ and the process is ‘generally employed, with self-evident symbolism, to indicate such concepts as distribution, plurality, repetition, customary activity, increase of size, added intensity, continuance’. Moravcsik (1978: 328) sees reduplication ‘a true universal of human language’ in the sense that all languages have it, even if they use it for different purposes.

The terms repetition and reduplication are not synonymous, even if they are often used that way and the processes result in similar, often highly expressive, imitative and intensifying interpretations. When the terms are separated, reduplication is seen as a morphological process where a word or its sub-part is iterated for a particular grammatical function; see Haspelmath (2002), Gil (2005), Katamba & Stonham (2006), Crystal (2008), Inkelas & Downing (2015), Stolz & Levkovych (2018). The iterated material must be phonologically or morpho-semantically characterizable as a segment, such as a syllable, root, stem, affix or word (Moravcsik 1978: 303f.). Full or total reduplication involves iterating material in full, while partial reduplication involves iterating it partially. The process has been associated with various inflectional and derivational functions that introduce imagery of doubling, multiplication, increase, intensification, recursiveness or repetition. Reduplicatives are used to signal, among other things, plurality or distributivity of nouns, as in the Indonesian kanak ‘child’ vs. kanak kanak ‘children’ (Haspelmath 2002: 274); increased intensity of adjectives, as in the Turkish dolu ‘full’ vs. dopdolu ‘quite full’ (Moravcsik 1978: 308); and iterativity, progressivity or durativity of verbs, as in the Aztec woman ‘bark at’ vs. womwoman ‘he is barking at’ (Kay 1965, as cited in Moravcsik 1978: 308); see also Inkelas (2006), Katamba & Stonham (2006: 180–193), Booij (2007: 34–41), Aronoff & Fudeman (2011: 81, 175f.), Inkelas & Downing (2015), Lieber (2016: 92–94).

The term repetition is also used of processes that introduce imagery of intensification, increase, recursiveness or repetition. Depending on the researcher’s background and aims, such processes can be labelled imitation, echolalia, stuttering, cohesion, assonance or reiteration; see e.g. Watt (1968), Aitchison (1994:16, 21f.), Jackson (2016: 4f., 8–11), Stolz & Levkovych (2018).¹ The repeated material can again be a word or its sub-part, for example a morpheme or syllable as in the English (I am so) hap-hap-happy, but it can also be bigger than a word, as in (I am so) happy happy happy and Oh dear oh dear oh dear; see Aichison (1994), Travis (2001), Gomeshi et al. (2004), Gil (2005), Inkelas (2006), Jackson (2016), Stolz & Levkovych (2018). Although there is variation in the definitions and terminology used, most authors separate between a morphological process where the end-result is a single word(-sized unit), and a syntactic process where the end-result is a sequence of words or their parts, phrases or sentences. Another difference is that, although the processes result in sequences that may look similar, the sequences do not have the same functions and interpretations: only morphological processes can create expressions that are distinct lexical items. Syntactic processes create expressions that may be marked, intensified and emphasized, or where the repetition of material serves some other identifiable grammatical or semantic function, but there are no arguments for treating the sequences and their input material as separate lexical items (e.g. Gil 2005; Finkbeiner 2018; Stolz & Levkovych 2018). The size of the repeated material does not seem to affect these readings: both (I am so) hap-hap-happy and (I am so) happy happy happy convey extreme feelings of happiness.

Although there are problems with using word-hood as a criterion (Stolz & Levkovych 2018), the above considerations will serve as a starting point. The aim of the present paper is to (re-)evaluate the treatment of expressions like arty-farty, namby-pamby, hoity-toity, wishy-washy, lardy-dardy (Elton 1995–1996) as reduplicative or repetitive compound words in nearly all previous work on English. We examine, first, if such expressions meet the criteria for compounds and, second, if there are reasons to separate between expressions that may look similar, but which have different functions and interpretations and are unlikely to result from the same process/es. We begin by looking at accounts of English reduplicatives in section 2. We use terms like reduplicative, reduplication and repetition for convenience, without assuming anything (yet) about the status of such sequences. The focus will be on the grammatical functions; the status of such sequences as possible ideophones, in the sense of ‘marked words that depict sensory imagery’ (Dingemanse 2012: 655), is considered only marginally. We follow the spelling conventions in dictionaries, grammars and textbooks and spell most items with a hyphen, even if they are spelled in a variety of ways in real language examples and for many items, the spelling as two separate forms (e.g. teeny weeny, instead of teeny-weeny or teenyweeny) is the most popular alternative.

2. Reduplicative expressions in English: Previous accounts

In most English grammars and textbooks, reduplicatives are mentioned in passing, if at all, in connection with word-formation and compounding. Tenacious readers may be able to find information that is of relevance under labels like tautology, repetition, intensification, iteration, adverbs and interjections. While reduplicating material is seen as a productive process in English, it is more restricted than other word-formation processes. Reduplication is also a relatively recent process: according to Minkova (2002: 133f.), Old English texts contain occasional examples, and the main types were not established until the end of the 15th century.²

One of the earliest accounts of English reduplicatives is Wheatley’s (1866) Dictionary of Reduplicated Words in the English Language which lists over 700 partially reduplicative forms that are divided into (i) items that involve a change in initial consonant (fol-lol, namby-pamby, roley-poley); (ii) items that involve a vowel change (bibble-babble, chit-chat, flim-flam); and (iii) items where something is added to the second part (argle-bargle). Full reduplicatives are excluded from the dictionary, on grounds that they are present only ‘in the most uncultivated tongues’ (Wheatley 1866: 3). In most items, Wheatley sees the first part as ‘a mere prefix, used only for the purpose of strengthening or intensifying the meaning’, while the second part is the proper word (1866: 5). Examples include chit-chat (where chat is the proper word), dilly-dally (from dally) and rimble-ramble (ramble). Examples where the first part is the proper word include argle-bargle, handy-pandy (from hand), rowdy-dowdy (rowdy), and whim-wham (whim). Wheatley (1866: 7) observes that existing words can sometimes be combined solely on grounds that they rhyme: clap-trap, rumble-tumble and willy-nilly are cases in point. In terms of meaning, Wheatley assumes most reduplicatives to be ‘of an onomatopoetic character’ and ‘purely imitative of the sound or meaning intended to be represented’ (1866: 7). The remaining items are ‘meaningless words which are used in rhymes for chorusses’ and ‘comical perversions of correct words or terms’ (Wheatley 1866: 7f.).

Zandvoort’s (1948: 308–310) grammar covers both full and partial reduplicatives under the label repetition compounds. Most full repetition compounds, including drip-drip, pooh-pooh, clank-clank, thump-thump, are argued to ‘imitate or suggest (repeated) sounds’ (Zandvoort 1948: 308). The same holds for most partial repetition compounds, which are divided into the ablaut (dilly-dally, fiddle-faddle, flim-flam) and rhyming types (hob-nob, namby-pamby, hurly-burly from Zandvoort 1948: 309f.). Many of the items are seen as ‘nursery words’ and ‘nonce words’ (Zandvoort 1948: 310). This is a typical example of how reduplicatives are treated in grammars: the account consists of lists of items that are said to describe sound or identify entities producing the sound. The position is not questioned even when the items clearly do not describe sound, such as Zandvoort’s dilly-dally, fiddle-faddle, flim-flam, namby-pamby, hurly-burly. A similar approach is adopted in recent grammars. Huddleston & Pullum (2002: 1666) cover what they call phonologically motivated compounds in just half a page, even if relevant information can be located, in bits and pieces, in various other chapters as well. The account consists of lists of partial reduplicatives, including gang-bang, easy-peasy, helter-skelter, hoity-toity, chitchat, shillyshally, clipclop, dingdong. There is no discussion about the range of meanings such items can have and why the items are labelled compound words. Bauer’s (1983: 212f.) textbook on English word-formation also covers only partial reduplicatives, in a chapter on compound words. The labels ablaut and rhyme-motivated compound suggest that Bauer also sees such items as being phonologically motivated. There is no discussion of the range of meanings that such compounds can have, even if the examples cover a variety of them and items like stungun even intuitively feel different from items like hoity-toity. In their textbook on word-formation, Minkova & Stockwell (2009) argue that reduplication is both historically and in present-day English a ‘rather unimportant, though often amusing, resource for expanding the vocabulary’. The authors list many of the same examples, including hoity-toity, helter-skelter, flim-flam and shilly-shally, and propose that only a few of them ‘are more than trivial expansions of the vocabulary’ (2009: 22).

Although Quirk et al.’s (1972; 1985) grammars also cover reduplicatives in a few paragraphs, in connection with word-formation and compounding, the accounts differ from the others in that the items – this time labelled reduplicative compounds – are not limited to expressing sound. Quirk et al. (1972: 1029f.; 1985: 1579f.) list examples like goody-goody, tick-tock, flip-flop, dilly-dally, shilly-shally and teeny-weeny and propose that, although many of them mimic sound (tick-tock), they can also be used to ‘suggest alternating movements’ (flip-flop); ‘disparage by suggesting instability, nonsense, insincerity and vacillation’ (dilly-dally, shilly-shally); and ‘intensify’ (teeny-weeny, tip-top). Quirk et al. do not motivate the compound status of the items and see them as mainly belonging ‘to the sphere of child-parent talk’ (1985: 1579).

Jespersen’s (1942: chapter X) grammar is one of the few sources where the resemblance between form and meaning is considered in more detail. Jespersen treats reduplicatives as a sub-type of compounds and such formations are, in contrast to Minkova & Stockwell (2009), seen as ‘very frequent in English’ (1942: 174). Fully reduplicative compounds, including chip-chip, drip-drip, gobble-gobble, quack-quack, thump-thump, are argued to express ‘repeated sounds, natural or produced by human activity’ (1942: 174). Partially reduplicative compounds are divided into the vowel change (ablaut) and consonant change (rhyming, but see below) sub-types. Examples of the former include chit-chat, clitter-clatter, dilly-dally, gibble-gabble, drip-drop, sing-song, tip-top. Their function, Jespersen (1942: 176) argues, is to ‘express the sound produced by the movement to and from or the movement itself as a zig-zag, hence vacillation, indecision, etc., and contemptile things in general’. Such readings are attributed to the form of the compounds: alternation between high and low vowels is seen as sound-symbolism that introduces associations with imagery of back and forth movement, in both concrete and metaphorical sense (Jespersen 1942: 176). A similar position is adopted in Marchand (1960: 347f.) who sees the vowel alternation as signalling ‘polarity’ in domains that range from imagery of concrete sound and movement to extended meanings like undecidedness, hesitancy and double-faced character. Forms where the initial consonant changes – namby-pamby, hodge-podge, teeny-weeny, ragbag, hoity-toity, lardy-dardy and so on – are argued to have ‘as a rule a less serious character’ and mainly ‘belong to the nursery’ (Jespersen 1942: 180). Marchand (1960: 345) agrees that, although both ablaut and rhyming reduplicatives are playful in tone, the former are so ‘in a neutrally aesthetic way’, the latter ‘in a childish, even babyish manner’. It should be noted that Jespersen sees the change of initial consonants as the main event, and the fact that the sequences also rhyme is secondary. Most other authors, including Marchand (1960: 348), view the rhyme as ‘obviously the basic factor in these combinations’.

In sum, there is agreement in previous work that English reduplicative compounds fall in three main sub-types, illustrated in Table 1 with examples from the sources just reviewed. There are consistent attempts to link each sub-type with a meaning: full reduplication is said to introduce associations with imagery of repetitive sound or movement, while vowel alternation in ablaut reduplicatives signals properties like alternation, vacillation, indecision and ‘contemptile things in general’ (Jespersen 1942: 176). Another common denominator is the view of reduplicatives as compound words, even if some sources admit that they are atypical compounds. There are no attempts to determine what criteria for compounds reduplicatives meet, and no attempts to separate possible compounds from other iterative sequences.

Table 1. Reduplicative types in English.

Full reduplicative compounds	goody-goody, pretty-pretty, girly-girly, clink-clink, clank-clank, drip-drip, thump-thump, […]
Alliterating/ablaut reduplicative compounds	clink-clank, drip-drop, flip-flop, teeny-tiny, tip-top, dilly-dally, shilly-shally, wiggle-waggle, pitter-patter, wishy-washy, chitchat, fiddle-faddle, knick-knack, mishmash, tittle-tattle, pingpong, singsong, […]
Rhyming reduplicative compounds	hoity-toity, helter-skelter, hodge-podge, hotchpotch, hobnob, ragbag, teeny-weeny, easy-peasy, willy-nilly, lovey-dovey, airy-fairy, lardy-dardy, arty-farty, razzle-dazzle, rumble-tumble, super-duper, claptrap, hanky-panky, […]

3. Compounds or not: A first look

All major English grammars – Quirk et al. (1972; 1985), Biber et al. (1999), Huddleston & Pullum (2002); see also Marchand (1960), Thun (1963), Fabb (1998), Bauer (2001), Plag (2003), Katamba & Stonham (2006), Lieber & Stekauer (2009), Bauer et al. (2013), Bauer (2017) – view compound words as lexical units that are constructed of other lexical units (i.e. roots, stems, bases, words, or free forms, depending on what terms the author prefers), at least one of which carries a meaning. In most compounds, the constituent parts are independent words that can stand on their own and occur in other combinations. The meaning of the whole differs from the meanings of one or both of the constituents: hatpin does not denote the same object as hat or pin. Most English compounds are noun-noun combinations that are headed by the rightmost constituent, which means that possible inflection is usually visible on the rightmost constituent: the plural of hatpin is hatpins and the plural of housemouse is housemice. Compound words are internally stable: the order of constituents is fixed and there are no alternative variants, apart from those resulting from normal derivational and inflectional processes. Material cannot be inserted freely between the constituents, the constituents cannot be iterated freely further, and they cannot be combined freely with other items that may or may not alliterate or rhyme with them and/or belong to the same semantic field. The keyword is freely: English has compounds where the constituents are joined by coordinators (hit-and-run, do-or-die, bread-and-butter), but these are part of the expression, rather than something speakers add at will. Compounds are also modified as units, and their constituent parts cannot be modified independently. Hence, green cannot be modified by degree adverbs, to create #very green tea, and colourless green tea is not a contradiction; see also Bauer et al. (2013).

Although there are issues with most definitions of compound words – as there are issues with most definitions of words (Bauer 1998; Bauer et al. 2013) – the above criteria allow us to assess the status of English reduplicatives as possible compound words. In sections 3.1 and 3.2, we look at items that are the ‘easiest’ to explain from this perspective.

3.1. False reduplicatives that are real compounds

Items like cookbook are sometimes listed as examples of reduplicative compounds. Thun (1963) and Bauer et al. (2013) are among the few authors who separate cookbook, down-town, ill-will, night-light, mole-hole, housemouse, may-day, pay-day from real reduplicatives. Thun calls these false reduplicatives and argues that the rhyme – according to Thun they always rhyme – is a coincidence. Ill-will and night-light are, in Thun’s (1963: 14) view, formed after their opposites goodwill and daylight. Mole-hole is part of an existing pattern alongside rat-holes and mouse-holes: it is a coincidence that a hole made by a mole constitutes a rhyming pair. This line of reasoning cannot explain items like down-town, as etymological dictionaries show that the item, originally a phrase down-in-the-town, pre-dates its opposite uptown (OED s.v. downtown). Thun’s list also contains items that lack opposites, for example cookbook, which is based on the earlier form cookery book (OED s.v. cookbook). If phrases can be reduced to rhyming pairs and suffixes are dropped for the same effect, one may wonder how accidental the rhyme really is. Terms like phonologically motivated (Huddleston & Pullum 2002: 1666) and ablaut/rhyme-motivated (Bauer 1983: 212f.; see also Katamba & Stonham 2006: 305f.; Bauer et al. 2013) suggest that the alliteration or rhyme may in fact drive word-formation and change. A similar position is adopted in Benczes (2012; 2014), who argues that speakers often prefer, especially in informal domains, combinations that are alliterating or rhyming pairs; see also Bolinger (1950).

Despite the alliteration or rhyme being possible driving forces behind word-formation and change, there are reasons to treat cookbook and so on as real compounds that are not real reduplicatives. First, they are combinations of independently attested words whose meanings are based on the meanings of the constituents. Neither of the constituents is formed on the basis of the other, and their meanings are not related – a point that will become clearer in section 5. Secondly, they are mainly nouns, rather than multicategorial – another point that will become clearer shortly. Third, the order of their constituent parts is fixed, and there are no alternative variants: cookbook can only alternate with the earlier form cookery book. The constituents cannot be iterated freely further, and their form – the fact that they alliterate or rhyme – is not associated with an identifiable function, such as creating imagery of increase, intensification, recursiveness or repetition of something. Finally, the cookbook words are not restricted to informal domains, the way items like arty-farty, namby-pamby, hoity-toity often are.

The above criteria also capture items like toyboy, bigwig, cop-shop, flower-power, shitfit. While their meanings are less transparent than in the cookbook set, they are constructed of distinct words and their meanings are, despite being playful, still compositional. They are mainly nouns, the order of their constituents is fixed, there are no alternative variants, and material cannot be added or iterated freely further. Although the alliteration or rhyme may be an even stronger reason for creating these items (Bauer et al. 2013) and they are mainly used in informal domains, their form is not associated with a distinct function, such as associations with imagery of increase, intensification, recursiveness or repetition of something in the world. These observations suggest that English has compound words whose constituents alliterate or rhyme. Although the alliteration or rhyme can be accidental, many of these compounds are likely to be formed because of the alliteration or rhyme. As we soon see, there are reasons to keep the compounds separate from most other reduplicatives in English.

3.2. Partial reduplicatives that are not compounds

Items like pell-mell, helter-skelter, higgledy-piggledy, topsy-turvy are common examples of reduplicative compounds in English; see e.g. Jespersen (1942: chapter X), Bauer (1983: 212f.), Quirk et al. (1985: 1580), Huddleston & Pullum (2002: 1666). Labelling them compound words is problematic, as the items are not constructed of meaningful component parts. Many of these items have been borrowed into English in a seemingly already compounded form: pell-mell originates from French, with the etymons pelle-melle or pêle-mêle (OED s.v. pell-mell). It is seen as a synonym to helter-skelter, which is a ‘jingling expression vaguely imitating the hurried clatter of feet rapidly and irregularly moved, or of many running feet’ (OED s.v. helter-skelter). There is no mention of helter-skelter having been borrowed or formed of helter or skelter, neither of which is recognized as a meaningful item. Similar observations hold for higgledy-piggledy, which is labelled a ‘rhyming compound of obscure origin’ and which, contrary to popular belief, is not based on pig (OED s.v. higgledy-piggledy). Topsy-turvy is described as an ‘alliterative or assonant combination’ and the OED lists several possible etymologies, none of which is more certain than the others (s.v. topsy-turvy). Crucially, neither topsy nor turvy is assumed to be, or to ever have been, a meaningful word in English.

Although there is no doubt that these items would not be meaningful words in present-day English, they cannot be compound words, on grounds that they are not divisible into meaningful constituents – that they may have been compound words in some other language is not an argument for treating them as such in English. The items can, of course, be divided into constituents that alliterate or rhyme, but the constituents are not associated with a meaning any more than, say, syllables are. The alternative is to treat the pell-mell items as words in their own right. Like other word(-sized unit)s, they are internally stable: material comes in fixed order and cannot be iterated freely further. Their form – the alliteration or rhyme – is not associated with a grammatical or semantic function, such as associations with imagery of intensification, increase, recursiveness or repetition of something in the world. A further observation is that, unlike the items in section 4.1, most pell-mell items are multicategorial and are defined in dictionaries as adverbs, adjectives, nouns and verbs, sometimes also interjections.

The pell-mell words overlap with items that are, originally, composed of independent words but which may no longer be recognized as such. Harum-scarum is constructed of the verbs hare ‘to run or move with great speed’ (which is derived from the noun hare; OED s.v. hare) and scare ‘to frighten, terrify’ (OED s.v. scare). This suggests that harum-scarum has started out in the cookbook/toyboy set. Although its current meaning ‘Reckless action or behaviour’ (OED s.v. harum-scarum) may be linked with imagery of a scared hare, the fact that present-day speakers are probably unaware of this speaks in favour of treating it as a lexical item that is not constructed of other lexical items. Most harum-scarum items are multicategorial, the order of constituents is fixed, there are no alternative variants, and material cannot be inserted or iterated further. Their form is not associated with an identifiable meaning or function, such as adding associations with imagery of increase, intensification, recursiveness or repetition, even if they are often playful in tone and are mainly restricted to informal domains. Items like hoity-toity, argle-bargle and namby-pamby may also be placed in this set. Hoity-toity is ‘apparently a derivative of hoit v., with reduplication’ (OED s.v. hoity-toity), which means that one of its constituents has, once upon a time, been a distinct lexical item. As hoit is no longer recognized as such, the meaning of hoity-toity cannot be linked to the meaning of either constituent. The same holds for argle-bargle, which is constructed of the now obsolete word argle (OED s.v. argle-bargle). Items constructed of proper nouns whose referents are no longer known may also qualify as members of this set. Namby-pamby was originally a derogatory nickname for the 18th century poet Ambrose Phillips: in present-day English, neither namby nor pamby is associated with a referent or meaning that explains why the word denotes entities ‘inclined to weak sentimentality, affectedly dainty; lacking vigour or drive’ (OED s.v. namby-pamby). Hoity-toity, argle-bargle and namby-pamby display the same properties as discussed above: they are multicategorial and internally stable, they lack alternative forms, material cannot be inserted or iterated further, and their form does not serve an identifiable grammatical or semantic function.

The conclusion for the expressions in this section is that they are certainly words, but not compound words. A common denominator of the real compounds in section 3.1 and the items discussed here is that their forms are stable, which means that they do not usually overlap with other sequences where material is iterated. This distinguishes them from most other alleged reduplicative compounds, to be discussed in sections 4 and 5.

4. Full reduplicatives and other fully iterative sequences

English has two main types of alleged fully reduplicative compounds: items that are formed of (seemingly) referential words and items that are formed of expressive/onomatopoeic words. We start by looking at the first type in section 4.1.

4.1. Entities that are pretty-pretty need not be pretty pretty pretty

Items like pretty-pretty, goody-goody, girly-girly, preachy-preachy are listed as examples of fully reduplicative compounds in e.g. Jespersen (1942: chapter X), Zandvoort (1948: 308f.), Thun (1963: 145), Quirk et al. (1972: 1029f.; 1985: 1579f.) and Katamba & Stonham (2006: 180f.). They are mainly adjectives and nouns, and the meaning of the alleged compound is more or less the same as the meaning of its sole constituent. The pair pretty-pretty vs. pretty is exceptional, in that the meaning of the reduplicative differs from the meaning of the single form. The definitions in Table 2 reveal that the reduplicative is associated with such excessive prettiness that it reverses the original meaning of pretty and becomes mockery.

Table 2. The meaning explanations of the adjectives pretty-pretty and pretty.

pretty-pretty, adjective	‘excessively, cloyingly, or affectedly pretty or sentimental’ (OED s.v. pretty-pretty)
pretty, adjective	‘clever, skilful, able’/‘Cleverly or elegantly made or done; ingenious, artful, well-conceived’/‘attractive and pleasing in appearance; good-looking, esp. in a delicate or diminutive way’/‘pleasing to the senses, aesthetically pleasing; attractive or charming’ (OED s.v. pretty)

Present-day speakers do use pretty-pretty, in its various spelled forms, to signal the meanings outlined in Table 2. When used with this interpretation, pretty-pretty can cooccur with negative words, such as the adjective meaningless. The examples in (1) come from a standardized corpus; the relevant words are in marked in bold in all examples.³

Table

(1)	a.	She sings it in a meaningless pretty-pretty way, and […] (iWEB)
	b.	Flaubert who loathed pretty-pretty prose would have applauded Kafka’s attitude […] (iWEB)

A closer look at naturally occurring examples reveals, however, that most occurrences of pretty-pretty, irrespective of the spelling, are examples of something else. One function is so-called contrastive or SALAD-salad reduplication, based on a famous example by Gomeshi et al. (2004); see also Travis (2001), Bauer et al. (2013), Bauer (2017). Contrastive reduplication can target words, phrases and sentences. Material is repeated exactly once, and the resulting items denote prototypical instances of their referents; see (2).

Table

(2)	a.	She wasn’t pretty-pretty – he had been with girls that were pretty-pretty – but she was pretty (COCA 2019, FIC)
	b.	I think she’s pretty. But she’s not, like pretty-pretty. (COCA 2016, TV)

Other functions include the use of pretty-pretty as normal adjectival modifier of nouns (3) and degree modifier of adjectives (4). In both cases, iteration introduces an intensifying reading that does not reverse the meaning of pretty. The sequences can be replaced by (sequences of) degree adverbs: (3a) can be paraphrased as very pretty shelves, (4a) as very very good.

Table

(3)	a.	Those will stay on my shelves, my pretty-pretty shelves, […] (COCA 2012, MAG)
	b.	inside all those boxes are my pretty pretty buttons and beads collected over the years. (iWEB)
(4)	a.	How’s that sound? Yeah, pretty-pretty good. (COCA 2014, MOV)
	b.	[…] there is a pretty pretty good chance he’s into that you too. (COCA 2012, BLOG)

In (3)–(4), speakers can repeat pretty any number of times. Watt (1968) had argued that further iteration does not imply a further increase in intensity, which means that (3)–(4) are not less intensifying or involve less prettiness than (5).

Table

(5)	a.	Since then, i have had lots of pretty pretty pretty girls […] (GloWbE, KE G)
	b.	I suddenly became very interested in making my home pretty pretty pretty, but […] (COCA 2012, BLOG)

As iteration serves a function, we expect (3)–(5) to be incompatible with expressions that deny this function, that is, deny the increased intensity or amount. The contrast between (6a, b) which are based on (5b) shows that the expectation is met.

Table

(6)	a.	I made my home pretty, but not that pretty.
	b.	#I made my home pretty pretty pretty, but not that pretty.

Goody-goody, girly-girly and so on are based on constituents that are recognizable as derived words: the – y is an affix that adds meanings like ‘full of’ or ‘characterized by’ (Minkova & Stockwell 2009: 106; OED s.v. -y suffix).⁴ Definitions of goody-goody and goody show that both the reduplicative and the single form are associated with readings that are ‘Frequently in explicit contrast with good’ (OED s.v. goody-goody, goody). This reading is illustrated in (7)–(8), where the reversal or denial of the meaning of good is strengthened further by negative words like valueless, lousy or uptight. Another observation is that the number of copies seems to be restricted: only goody, goody-goody, goody-good and the expression (goody) goody two-shoes, in their various spelled forms, have such interpretations.⁵

Table

(7)	a.	So all around me there’s these three girls that are so goodygoody they’ve probably never farted, […] (COCA 1997, FIC)
	b.	We can mistakenly use it as a disguise for valueless, goody-goody talk about anything made by anybody,[…] (COCA 2000, ACAD)
	c.	I can pay back those lousy goody goody Gryffindors. (iWEB)
	d.	There is nothing in the world as charming as watching the goody, goody uptight alpha discover his inner bad boy […] (iWEB)
(8)	a.	how is it that these flabby philanthropists, these goody-good sloppy sentimentalists invariably champion […] (iWEB)
	b.	[…] which meant weird and immature, nerdy, goody good Christian people. (iWEB)
	c.	They’re not Goody Two-Shoes type good. (iWEB)

A look at naturally occurring examples suggests that readings as in (7)–(8) are in the minority, and most sequences of goody are used to signal increased intensity or amount. The examples in (9) are associated with the original meaning of good, which is the last item in the sequence. The -y is not an affix, but instead a meaningless ‘extension’ (Jespersen 1942: 175) or ‘link syllable’ (Thun 1963: 222–225) that is added for ease of pronunciation. The sequences are paraphrasable with (sequences of) degree modifiers, for example very very good, and goody can appear a number of times, without implications of a further increase in intensity or amount.

Table

(9)	a.	The movie is nice and has goody good feeling as shown and depicted in the trailer […] (NOW)
	b.	It’s all about weight and being goody goody goody good? (COCA 2012, WEB)

(10) show that the sequences are incompatible with expressions that deny the increase in intensity or amount.

Table

(10)	a.	The film was good, but not that good
	b.	#The film was (goody) goody good, but not that good.

In sum, expressions that look similar may be linked with different functions and interpretations: there are expressions where the number of copies is restricted and the reduplicatives – and in many cases their sole constituents – are associated with a meaning that reverses, denies or mocks the original meaning of the input item. There are also expressions where the number of copies is free and where iteration introduces associations with imagery of increase or intensification of something in the world. The former may qualify as compound words; the fact that the compounds and their constituents mean more or less the same suggests that they have been idiomaticized and lexicalized. The latter appear to be iterative sequences of words, created by whatever processes are responsible for iterating or stacking words in English. We return to such processes briefly. In section 4.2, we look at the second – and by far the more common – type of alleged fully reduplicative compounds, namely items whose constituents are seen as expressive/onomatopoeic words.

4.2. Entities that clink-clink can also clink clink clink

Most clink-clink type expressions, as noted in section 2, are argued to be onomatopoeic and their forms are said to create associations with imagery of the sound and/or movement denoted by the sole constituent as being increased, intensified, repeated or continued; see e.g. Jespersen (1942: chapter X), Zandvoort (1948: 309f.), Thun (1963), Quirk et al. (1972: 1029f.; 1985: 1579f.), Huddleston & Pullum (2002: 1666). A property that has received little attention is that, unlike the items in section 4.1, most of these items are associated with only one reading: the reduplicative forms are associated with imagery of increased intensity, amount, repetition or duration, but otherwise the meanings are the same as the meanings of the input items. Some examples are given in Table 3; the -ety is another example of a link syllable (Jespersen 1942: 175; Thun 1963: 222–225).

Table 3. The meaning explanations of full reduplicatives based on expressive words.

clink(ety)-clink, noun	‘a repeated usually rhythmic clinking sound’ (Merriam-Webster s.v. clinkety-clink)
clink, noun	‘a clinking sound’ (Merriam-Webster s.v. clink)
drip-drip, noun	‘a continued dripping’ (Merriam-Webster s.v. drip-drip)
ticktick, noun	‘a repeated ticking sound’ (Merriam-Webster s.v. ticktick)

Another common property of these reduplicatives is that the number of copies is free: while we can distinguish between one and many, we cannot say that two copies would imply two clinks, any more than four copies imply four clinks (Watt 1968). Nor are the events in (11)–(12) viewed as being considerably shorter or less repetitive than in (13)–(14). All the examples imply sequences of events and are paraphrasable with expressions like nervous clinking, dismal dripping and clink-clink-and-yet-another-clink.

Table

(11)	a.	They were quietly discussing, over the nervous clink-clink of teacups, […] (iWEB)
	b.	[…] like some fabulous cloak unfurled against the dismal dripdrip of the steamers working up above. (COCA 1999, FIC)
	c.	I finish loading just as I hear a quiet thump-thump from the other room, and […] (GloWbE, GB G)
(12)	a.	The bell clink-clinked as they walked out the door hand in hand, […] (COCA 2017, FIC)
	b.	I thought you might drip-drip sweat into the tea. (COCA 2012, BLOG)
	c.	[…] as her coach thump-thumped over a familiar wooden bridge. (GloWbE, TZ G)
(13)	a.	[…] they announce your presence with a distinct clink-clink-clink that can get aggravating, fast (iWEB)
	b.	This atmospheric river coupled with the drip, drip, drip of snow melting in […]. (Clements 2022)
	c.	Minutes later, the group heard the thump-thump-thump of the choppers blades […] (iWEB)
(14)	a.	Rye is not just a type of whiskey; it’s the whiskey that the gunslingers demanded when their spurs clink-clink-clinked into the saloon. (When you drink rye, you are drinking American history n.d)
	b.	The blood dripped down and onto my lips and I eagerly lapped up the dropping blood as it drip, drip dripped into my mouth (iWEB)
	c.	[…] the bumper of a passing auto as its tires thump, thump, thumped across the wooden cross beams. (Lahurd n.d)

The main reason why material is reduplicated seems to be to signal an increase or repetition of something in the world: a case of diagrammatic iconicity. Watt’s (1968) observation that further iteration does not seem to signal further repetition is worth bearing in mind in the later sections. Most clink-clink type items are multicategorial and are used as adverbs, adjectives, nouns and verbs, sometimes interjections; this holds of the single forms as well. Only a handful of expressions are more similar to pretty-pretty and goody-goody in section 4.1 and have two functions. The compound-like functions and interpretations appear to be metaphorical extensions, where the associations with imagery of sound and movement have been obscured. Both bling-bling and bling, exemplified in (15), can be used to denote ‘flashy jewelry worn especially as an indication of wealth or status’ (Merriam-Webster s.v. bling-bling, bling). The sound imagery is still present in definitions how ‘light glinting off gold or diamonds would make a bling-like sound if you could hear it’ (Vocabulary.com s.v. bling). Most examples of bling-bling and bling that can be located in standardized corpora refer to jewelry, clothing or a lifestyle. The number of blings is restricted to one or two, although there are occasional examples of three or more blings being used this way. (16) exemplify the (rare) cases where sequences of bling depict repeated continuous sound, real or imaginary, associated mainly with social media and/or mobile phones.

Table

(15)	a.	As far as I’m concerned, the more crowns, tiaras and general bling the merrier, […] (NOW 23–02-16, GB)
	b.	It isn’t about luxury and bling-bling. (NOW 22–03-15 GB)
(16)	a.	I woke up and of course social media was bling bling bling all over the place. (NOW 18–11-05 US)
	b.	It was like bling bling, bling bling, bling, bling, bling, bling. My phone was blowing up […] (NOW 22–10-07 US)

The fact that sequences of expressive/onomatopoeic words are used to signal primarily intensifying, repetitive or continuative readings and the number of copies is free suggests that the sequences are not compounds words. The fact that clink, drip and thump can be used as verbs, and verbs display more overt inflection in English than nouns and adjectives do, raises further questions about their status. There are, on the one hand, sequences like (12) and (14), where the rightmost item is overtly inflected and, on the other hand, sequences like (17), where all items are inflected. Sequences like (17) overlap with sequences like (18), where the items are conjoined by overt coordinators.

Table

(17)	a.	We listened to foghorns and sea lions in the distance, watched Pelicans dive-bomb fish in the harbor, counted dew drops as they dripped dripped dripped onto the boat, […] (Erin n.d)
	b.	Her heart thumped thumped thumped hard in her chest. (Smith 2017)
(18)	a.	I asked Jimmy to fix it every day, and every day it just dripped and dripped. (COCA 2013, FIC)
	b.	So he gave the pitcher a hard thump. It did not break. He thumped and thumped and thumped, first here, then there. (iWEB)

All these examples are associated with imagery of increase, repetition and continuity of the process or event denoted by the verb. This imagery is a property of the sequence, the individual members retain their original meanings, and the number of copies is free. Yet (17)–(18) do not appear to mean exactly the same as (12) and (14). Although the differences are subtle, speakers report that the events in (17)–(18) feel bigger and longer-lasting than the ones in (12) and (14). As one native speaker put it, blood that dripped dripped dripped gives the impression of a heavier bleeding that ‘just won’t stop’. Another difference is that an event where something clink-clink-clinks is felt to consist of sub-events of the same type. An event where something clinks clinks clinks or clinks and clinks and clinks may be interpreted so, but it may also be understood so that each clink is a complete event, without sub-events. Some of these intuitions can be captured with tests that measure amount and duration (Watt 1968). The decreased acceptability of (19b), (20b) is expected, if we are dealing with a heavier bleeding that takes longer to stop.

Table

(19)	a.	Blood drip-drip-dripped from the wound, but not severely.
	b.	#Blood dripped dripped dripped from the wound, but not severely.
(20)	a.	Blood drip-drip-dripped from the wound for three minutes/for several hours.
	b.	Blood dripped dripped dripped from the wound #for three minutes/^okfor several hours.

Examples of sequences where only the last item is inflected and where all items are inflected are provided in e.g. Jespersen (1942: 173, 177), who does not assume any meaning difference between such formations. Thun (1963: 24) observes that ‘most readers will take pit, pit, pit to be three words’ whereas ‘few are likely to regard pit-pit-pitting as three words’ but he does not comment on a possible meaning difference between such formations. The same applies to items that are joined by overt coordinators: Jespersen (1942: chapter X) lists such items but views the material in between as an ‘intercalation’ that is similar to a link syllable. Thun (1963: 27) attempts to separate between real reduplicatives and ‘phrasal reduplicatives’ that are joined by and, but admits that the distinction is difficult to make. Thun (1963: chapter 15) lists numerous examples where the constituents are separated by intervening material, and the assumption seems to be that the readings are unaffected by such material. To be able to account for the facts we have just observed, we need to take a small detour and look at sequences of ‘ordinary’ words. The aim is to see if the accounts proposed for them can be extended to the data discussed in the present section.

4.3. An awful mess mess mess: Other fully iterative sequences in English

All major English grammars – Quirk et al. (1972; 1985), Biber et al. (1999), Huddleston & Pullum (2002); see also Watt (1968), Jackson (2016) – discuss examples where ‘ordinary’ words are iterated or stacked.⁶ There are no attempts to link these to the types of sequences in sections 4.1 and 4.2. (21) exemplify nouns, (22) adjectives. In both cases, iteration signals an increase in intensity or amount. This reading is a property of the sequence, while the individual members retain their original readings.

Table

(21)	a.	[…] oil in filters, misaligned spark plug wires, dirt/mud in the brake pads, dirt dirt dirt everywhere. (iWEB)
	b.	[…] because all I want to do is have cake cake cake!!! (iWEB)
(22)	a.	[…] remember coming home at the end of the day red, hot but cold and sick sick sick (iWEB)
	b.	If I tried to become a vegetarian I would be sick sick sick all the time. (iWEB)

As iteration is associated with a function, the sequences cannot combine with expressions that deny this function; see (23)–(24).

Table

(23)	a.	There was dirt, but not that much dirt.
	b.	#There was dirt dirt dirt, but not that much dirt.
(24)	a.	I was sick, but not that sick/not very often.
	b.	#I was sick sick sick, but not that sick/not very often.

(25)–(28) exemplify verbs. With state verbs, iteration adds associations with imagery of increased intensity or amount: how much one loves or hates. With dynamic verbs, iteration signals increased intensity or amount, or implies a repeated or continuous event. Although different lexical items behave differently, it is possible to find examples where inflectional affixes are realized on the rightmost item, or all items.

Table

(25)	a.	That boy love love lOVEs to Swish n Swipe, see him go!!!! (iWEB)
	b.	I had love love loved the other books in the series and […] (COCA 2012, WEB)
	c.	I’ve loved loved loved scrapbooking and paper crafting since I was a little girl […] (COCA 2012, BLOG)
(26)	a.	[…] I hate hate hated deceiving people. (COCA 2012, FIC)
	b.	[…] but I’ve hate hate HATED that picture for a very long time, so […] (GloWBE, GB G)
	c.	Tonight I saw “Star Wars: The Special Edition “and I hated hated hated it! (COCA 1997, NEWS)
(27)	a.	[…] they got no exercise, they just eat eat eat, that’s all they do. (COCA 2001, ACAD)
	b.	[…] and 2 rats have been eating eating the same hunk of flesh for the last 10 minutes – sigh! (iWEB)
	c.	He’s always eating eating eating, and in my mind i’m like ugh gross. (iWEB)
(28)	a.	Let us know why youre sad and cry cry cry and tell us why while you keep crying! (iWEB)
	b.	u think the puppy will become blind? it hurt me to heard the puppy crying crying.. (iWEB)
	c.	If she gets a little scratch she cries cries cries. (shay-doe 2020)

The sequences are incompatible with expressions that deny the intensified readings; see (29)–(30). This seems to be the primary function, as associations with continuity and durativity can, surprisingly, be denied. The examples of state verbs in (30) are marginally acceptable, with a reading where the intense sensations of love or hatred do not last long, on grounds that state verbs are assumed to denote long-lasting states; the single forms are equally marginal, as shown by ??I loved scrapbooking, but not for very long. (31) illustrate the same contrasts.

Table

(29)	a.	I love love loved scrapbooking, but #not that much/??but not for very long.
	b.	I hated hated hated the film, but #not that much/??not for very long.
(30)	a.	I ate ate ate, but #not that much/^oknot for very long.
	b.	I cried cried cried, but #not that heavily/^oknot for very long.
(31)	a.	I love love loved scrapbooking #for three minutes/??for three days/^okfor three years.
	b.	I ate ate ate for three minutes/for three days/for three years.

Like clink, sequences of ‘ordinary’ verbs can be joined by overt coordinators. Sequences like (32)–(33) are incompatible with expressions that deny the increase in intensity or amount, but they can combine with vague time expressions; see (34)–(35). More specific time expressions are mainly possible with dynamic verbs, as shown in (36).

Table

(32)	a.	And believe me, gifted’ imperfect’ quilts are just loved and loved and loved some more by the lucky recipients! (iWEB)
	b.	a movie called The Thing, and it came out in 1982 and it was hated and hated by the fans. (iWEB)
(33)	a.	We ate and ate and ate and still couldn’t get enough of all the great food (iWEB)
	b.	They cried and cried and cried about it but I would not budge. (iWEB)
(34)	a.	I loved and loved and loved the quilt, but #not very much/??not for very long.
	b.	I hated and hated and hated the film, but #not very much/??not for very long.
(35)	a.	We ate and ate and ate, but #not very much/^oknot for very long.
	b.	They cried and cried and cried, but #not very heavily/^oknot for very long.
(36)	a.	I loved and loved and loved the quilt #for three minutes/??for three days/^okfor three years.
	b.	We ate and ate and ate for three minutes/for three days/for three years.

The fact that members of different lexical categories behave differently has been noted in grammars, but there is no single explanation that would cover all these cases. In Huddleston & Pullum (2002: 561), sequences of adjectives are treated as examples of what is called intensificatory repetition. The process is said to have a similar effect to degree modification; this is precisely what we have seen in Section 4.1 and the present section. Watt (1968) has noted, furthermore, that repetition is primarily a property of adjectives that imply endpoints on a scale: good vs. bad, big vs. small, hot vs. cold. This may explain why The film was goody goody good is fine, while #The film was okay okay okay sounds odd. Quirk et al. (1985: 980f.; 1441) agree that sequences of adjectives signal an intensifying reading. Sequences of nouns convey, in their view, an idea of increased amount, which is certainly true of examples like (21a), while sequences of verbs imply repeated or continuous events, which is true primarily of dynamic verbs like (27)–(28). These analyses are in line with what we have observed for noun and verb sequences in section 4.2. The conclusion that suggests itself is that the non-compound sequences in sections 4.1 and 4.2 are further examples of the syntactic processes that iterate or stack ‘ordinary’ adjectives, nouns and verbs in English, and applying the term compound word to such sequences is misleading.

Looking at ‘ordinary’ words does not shed new light on the possible differences between sequences like clink-clink-clinked, clinked clinked clinked and clinked and clinked and clinked, however. Quirk et al. (1985: 980f.) mention such sequences in passing and propose that having the coordinator present is the preferred option. In Huddleston & Pullum (2002: 1304f.), coordinated sequences are brought up under the terms intensifying and coordinator-marked reduplication. Such sequences are argued to signal ‘a large number or amount’ when the process is applied to nouns, and a ‘high degree of continuity or repetition’ when it is applied to verbs. All items must be separated by an overt coordinator, Huddleston & Pullum argue, and examples like *I laughed, laughed, and laughed are marked as ill-formed. Such examples do exist, however, (37) being cases in point.

Table

(37)	a.	I walked walked and walked religiously everyday […] (iWEB)
	b.	[…] and so I waited waited waited and waited praying for my bank account to miraculously explode. (iWEB)
	c.	OMG I woud have cried cried and cried some more!! (iWEB)

In sum, the parallels between the non-compound sequences in sections 4.1 and 4.2 and the sequences of ‘ordinary’ words discussed in this section suggest that none of these sequences are real compound words. Instead, they are sequences of words and/or their parts that have been iterated, for a particular semantic purpose.

5. Partial reduplicatives and other partially iterative sequences

English partially reduplicative compounds can be divided in two main types: items where both constituents are independent words, and items where only one of them is. We start with the first group and look at cases where the constituents are mainly adjectives.

5.1. Just a teeny-weeny, teeny tiny very weeny problem

Teeny-weeny, teeny-tiny, easy-peasy and tip-top are common examples of English partially reduplicative compounds – see e.g. Wheatley (1866), Jespersen (1942: chapter X), Quirk et al. (1972: 1029f.; 1985: 1579f.), Huddleston & Pullum (2002: 1666), Katamba (2005: 54). They are also discussed under terms like repetition/synonymous/co-synonymous/tautological compounds in previous work; see e.g. Renner (2008), Benczes (2014), Bauer (2017), Bonacchi (2017), González-Días (2021). Their constituents are treated as distinct words in present-day English: they are mainly adjectives that can be seen as referential, in identifying properties or qualities, and expressive/iconic, in that their form (the speech sounds they contain) is assumed to mimic the properties or qualities in question. Teeny and weeny contain high front vowels that represent smallness and lightness, while the unvoiced plosives in tip and top are seen as imitative of sharpness and pointiness; see e.g. Bolinger (1940; 1950), Diffloth (1994), Winter & Perlman (2021), McLean et al. (2023), Winter et al. (2023). Another property of these expressions – and what separates them from the real compounds in section 3.1 – is that, while the form of the constituents may change, their meanings are usually the same, to the point that one or both items are defined in terms of the other. The same is true of the reduplicatives, as shown in Table 4.

Table 4. The meaning explanations of teeny-weeny and teeny-tiny.

teeny-weeny, adjective	‘Extremely tiny; very small’ (OED s.v. teeny-weeny)
teeny-tiny, adjective	‘Extremely tiny; very small.’ (OED s.v. teeny-tiny)
teeny, adjective	‘Very tiny, very small’ (OED s.v. teeny)
weeny, adjective	‘Very small, tiny’ (OED s.v. weeny)
tiny, adjective	‘Very small, little, or slight; wee, minute’ (OED s.v. tiny)

The similarities in form and meaning often result from one of the constituents being based on the other. In section 2, we have seen how one of the constituents is a ‘playful prefix’ or ‘comical perversion of the correct word or term’ (Wheatley 1866: 7f.). Allowing playful prefixes to become words is part of normal word-formation: teeny is not exceptional in being a ‘variant or alteration’ of tiny, which is the original word (OED s.v. teeny). Weeny is formed of the dialectal word wee, by adding an ‘ending imitated from tiny adj., teeny adj.2’ (OED s.v. weeny); the ending is, in fact, the same affix that we find in pretty, girly, and so on (OED s.v. -y suffix). There are also cases where one of the constituents is a shortened form of the reduplicative that is, in turn, constructed of the constituent that is the original word. Easy-peasy is described as a ‘reduplication of easy adj.’ that means ‘extremely easy, very simple’ (OED s.v. easy-peasy). Peasy is a distinct lexical item that is a clipping and means ‘very simple, easy’ (OED s.v. peasy).

Most of these sequences display clear un-compound-like behaviour. The constituents can appear in either order, even if one order seems to be preferred. The constituents can appear in other constellations and combine with other items in the same semantic field, especially if these alliterate or rhyme with them. This is in harmony with Bolinger’s (1950: 128) observation how forms that are ‘encountered in the same area of greater or lesser specificity of meaning, and are also closely similar in form, […] are likely to exercise a kind of magnetic attraction one upon the other’. The constituents can be premodified individually, which is strong evidence for non-compound status. The fact that speakers can iterate material freely points in the same direction. Relevant examples are provided in (38)–(41). It is worth noting that searches in all the big corpora have not resulted in examples where the sequences would have other readings than those described above; this does not mean that such readings are not possible, but to confirm if they are or not, one would need to use a different method.

Table

(38)	a.	[…] if our whole known Universe is not just a tiny teeny spec in an infinite ocean of specs […] (GloWbE, US B)
	b.	[…] but did find all the tiny weeny cloths I crocheted for her, many moons ago. (iWEB)
	c.	if i was a dude and had a weeny tiny pecker, i would not […] (iWEB)
(39)	a.	HTiB systems are usually very small and come with cute little teeny-weeny speakers. (iWEB)
	b.	[…] she became a Nintendo Game Boy addict because she loved “the teeny weeny little game packs, […] (COCA 1996, MAG)
	c.	my only teenyweeny wee issue was with raoul who was […] (iWEB)
(40)	a.	As they grow from a teeny-teeny tiny size to just a really small size, they […] (iWEB)
	b.	I’m a teeny teeny tiny girl and I already eat 24/7 as is. (iWEB)
	c.	[…] but I can’t justify that teeny weeny tiny risk of my kids not having a mother! (iWEB)
	d.	I was one of them and so was Geri Halliwell, who’s teeny weeny weeny. (NOW 21–04-03 IE)
(41)	a.	[…] it leaves you with little tiny very small nugs that fly off the joint, […] (iWEB)
	b.	It’s like a bunch of shortstories with a tiny very tiny thread between them. (iWEB)
	c.	[…] and you’ll see an itsy bitsy super tiny very small dolphin (iWEB)
	d.	Is anyone else out there a teeny little tiny bit relieved that Christmas is over? (Penny [@instap3n] 2022)
	e.	Find the teeny weeny little tiny blue marker in the middle of the line of text […] (William 2012)

The above observations suggest that teeny-weeny, teeny-tiny and so on are not real compound words. As we are mainly dealing with adjectives that function as modifiers of nouns and can be paraphrased with (sequences of) degree adverbs – a person who is teeny weeny weeny can also be characterized as being very very weeny/small – an alternative would be to view them as further examples of Huddleston & Pullum’s (2002: 561) intensificatory repetition. Another option that suggests itself is to treat them as examples of intensificatory tautology, in the sense of adjectives ‘with identical or nearly identical meaning’ (Huddleston & Pullum 2002: 562). Intensificatory tautology is limited to a handful of semantic domains, including size, and expressions like great big and tiny little are listed as examples. Although such expressions are constructed of material that is etymologically unrelated, while the items discussed here are related, there are important similarities between them: crucially, a great big house need not be bigger in size than a house that is ‘only’ big. The alleged compound status of teeny-weeny or teeny-tiny cannot either be justified by saying that they are more established as units than great big. The 14-billion-word iWEB corpus has 832 examples of teeny-weeny, 6.257 examples of teeny-tiny and 13.531 examples of great big, in the various spelled forms; for all items, the spelling as two separate words is by far the most common option. This suggests that great big is actually more established as a unit than teeny-weeny or teeny-tiny are.

If teeny-weeny or teeny-tiny are not compound words, but instead sequences of words, we would expect iteration to be linked with a semantic function. While it is true that examples like (38)–(41) are highly marked and emphasized, expressions like teeny-weeny and teeny tiny weeny do not seem to imply that an entity would be considerably smaller in size than an entity that is ‘only’ tiny. Although this may seem contradictory to what we have said above, it falls within the scope of Watt’s (1968) observation how only adjectives implying endpoints on a scale can iterate. Teeny, tiny, weeny denote properties that fall outside the range of a scale where the endpoints are big vs. small: entities that are teeny are already smaller in size than entities that are small. Premodifying such adjectives with degree adverbs does not change the basic set-up: entities that are teeny or very teeny can be characterized as very very small, just like entities that are teeny-weeny or very teeny-weeny can be very very small. Iteration will add emphasis and make the expressions even more marked, but it does not add anything to their meaning. This is also in line with what we have observed in section 4.2 for sequences of clink, drip, thump: we can separate between one and many, and adding more clinks, drips or thumps will add emphasis and make the sequences more marked, but it will not signal an even further increase in the amount of clinking, dripping or thumping.

5.2. Things that go clink clank and clinkety clank in the night

In this section, we look at alleged partially reduplicative compounds that are also constructed of independent words. Most of them are alliterating pairs of expressive words and, like the items discussed in section 4.2, multicategorial. Many of these items have fully reduplicative counterparts: clink-clank vs. clink-clink, drip-drop vs. drip-drip, tick-tock vs. tick-tick (Jespersen 1942: chapter X), Marchand (1960: 347f.), Thun (1963: 16f.). Although the form of the constituents changes, their meanings are often similar, to the point that Thun (1963: 16) has proposed the labels coordinative or synonym compounds for them. The similarity in meaning can be observed in the definition of clink as ‘A sharp abrupt ringing sound, clearer and thinner than a clank, as of small metallic bodies or glasses struck together’ (OED s.v. clink). The meaning of the reduplicative is also similar to the meanings of the constituents, apart from the increased intensity, repetition and continuity and, in contrast to the fully reduplicative forms, alternation. The definitions of clink-clank as ‘A succession or alternation of clinking sounds’ (OED s.v. clink-clank) and drip-drop as ‘Continuous dripping with alternation of sound’ (OED s.v. drip-drop) are good examples of this. The reduplicative forms, as noted in section 4.2, tend to imply a series of events: (42) are unlikely to denote a single clink followed by a single clank.

Table

(42)	a.	There’s a clink-clank of pencils on stands as players hunch into their folders. (NOW 20–03-26 US)
	b.	[…] food advocates from Agroecology Exchange (AEX) and participants clink-clanked utensils in the day-long, […] (UPLB Perspective 2020)

The constituents can come in either order, although one of them – usually the order where the high vowel sound is uttered first – is preferred. The constituents can combine with other items that alliterate or rhyme, with or without link syllables; see (43). Material can also be iterated further: while examples like (44a) exist, the preferred option is to repeat the entire pair. This is illustrated in (44b–d).

Table

(43)	a.	[…] causing a clankety clinking noise that is very off putting. (iWEB)
	b.	You dare to come to me for a heart, do you? You clinking clanking, clattering collection of caliginous junk? (iWEB)
	c.	Argo, realising the battle is lost clippity clops over the barriers and […] (iWEB)
(44)	a.	More than a few mornings we’ve woken to that drip drip drop of rain pouring onto our roof […] (iWEB)
	b.	This is the thing that makes a clock go tick-tock, tick-tock, tick-tock, tick-tock. (iWEB)
	c.	And a black horse that went clip clop clip clop through Hyde Park. (iWEB)
	d.	I want to walk with a brisk clip clop, clip clop, clip clop in these, but when I actually walk I go slither, slither, slide (iWEB)

Possible inflectional endings can be realized on the rightmost item (45), or all items (46).

Table

(45)	a.	The Claymore’s front mast ticktocked like a metronome, […] (NOW 22–03-20 US)
	b.	I wan na feel your heart go tic-toc While the rain from your hair drip-drops […] (iWEB)
(46)	a.	The faucet that drips drips drops during the day goes unnoticed […] (iWEB)
	b.	and she flips flops after every conversation with the many advisors she’s surrounded herself with. (iWEB)

The constituents can be joined by coordinators, and they can combine with other items in the same semantic field that also alliterate or rhyme; see (47).

Table

(47)	a.	The bracers and bangles along his arm clank and clink with each step, […] (iWEB)
	b.	Chainmail clinked and clanked in the noiselessness. (iWEB)
	c.	I imagine the house filled with music as little girls and boys clinked and clunked their way through their mothers piano books. (iWEB)
	d.	[…] nicotine stains, ink splodges and other dripped or dropped substances […] (iWEB)
	e.	we’re rewound now, and we have a whole lot more ticking and tocking to do. (COCA 2015, TV)
	f.	[…] the old brass clock that ticked and tocked laboriously next to the astrolabe […] (COCA 2005, FIC)
	g.	I flipped and flopped and flipped again on […] (iWEB)

Expressions like wiggle-waggle, pitter-patter, fiddle-faddle display similar properties. A look at dictionary definitions reveals how the constituents are each other’s synonyms: wiggle is said to mean ‘To move to and fro or from side to side irregularly and lightly, to waggle; […]’ (OED s.v. wiggle), while waggle means ‘To move […] to and fro with short quick motions, or with a rapid undulation’ (OED s.v. waggle). The meaning of the reduplicative is similar to the meanings of the constituents: wiggle-waggle is seen as a ‘Reduplicated form combining wiggle v. and waggle v. […], emphasizing the alternation of movement’ (OED s.v. wiggle-waggle), while pitter-patter is a ‘Reduplication (with vowel variation) of patter v.2’ that means ‘To run with short light-sounding steps; to beat with a rapid alternation of light taps or pats, as rain; to palpitate’ (OED s.v. pitter-patter). Iteration seems to emphasize mainly alternation, as the constituents denote processes that are already iterative in nature. The similarities in form and meaning can again be attributed to similar etymologies: one constituent may be based on the other, or be a clipping of the reduplicative, which is in turn constructed of the constituent that is the original word. The constituents can come in either order, even if one order (usually the one where the high front vowel it uttered first) is preferred. Material can be iterated further, and a common option is again to repeat the entire pair; see (48).

Table

(48)	a.	what the fuck with the wiggle-waggle fuckin’ wiggle-waggle? (COCA 2006, TV)
	b.	At length there was a pitter patter, pitter patter, and some bits of mortar fell from […] (iWEB)

Inflectional affixes can be realized on the rightmost item (49), or all items (50).

Table

(49)	a.	Miss Lillian Chatlove Tattlewit gracefully gathered her skirts, and coquettishly wiggle-waggled toward the Steinway-Grand; […] (iWEB)
	b.	I pitter-pattered slowly down the steep hill […] (iWEB)
	c.	I have been fiddle faddling with this for months. (iWEB)
(50)	a.	He likes the front cover as the monkey’s head can be wiggled waggled but the interior […] (Brylina 2013)
	b.	[…] fitted 5 channel RC, fiddled-faddled adding detail and additional functions like […] (iWEB)

The constituents can also be separated by coordinators; (51).

Table

(51)	a.	She loves to roll on her back and wiggle and waggle until she scratches that itch. (iWEB)
	b.	[…] greeted by a very happy and active dog, who was wiggling and waggling his back end all over the office. (iWEB)
	c.	You hear the sounds of a soft rain pittering and pattering. Pittering and pattering. (Poey et al. 2022)
	d.	[…] missed a shot on a run inside, after fiddling-and-faddling around out front, […] (NOW 22–05-03 US)

Most expressions are associated with only one function and interpretation. A handful of items may have two functions and interpretations, where one of the meanings seems to be lexicalized so that the items qualify as possible compound words. Flip-flop can denote ‘A plastic or rubber sandal consisting of a flat sole and straps’ (OED s.v. flip-flop) or, as in (47 g), continuous alternating sound and movement. In the first case, it is a noun, and its form is fixed: each constituent occurs exactly once. In the latter case, it can also be a verb, and the individual constituents or the entire pair may be iterated further. The iteration serves the function of creating imagery of repetition and alternation. As before, further iteration does not imply a further increase in the amount of repetitions. All this speaks in favour of treating sequences like clink-clank as results of the same process that creates sequences like clink-clink, and the process is likely to be a syntactic one, rather than a word-formation process in the sense of real compounding.

5.3. All that dilly-dallying, dallying and dilling can be a problem

In this section, we look at expressions where only one of the constituents is a distinct word in present-day English. Many common examples of alliterating compounds in Wheatley (1866), Jespersen (1942: chapter X), Quirk et al. (1972: 1029f.; 1985: 1579f.), Huddleston & Pullum (2002: 1666), Katamba (2005: 54), including dilly-dally, wishy-washy, chitchat, rimble-ramble, flim-flam, fall in this set. Some rhyming compounds, including hurly-burly, may also be placed here. As noted above, many of the items in sections 5.1 and 5.2 may have started out in this set. A look at naturally occurring examples suggests that some expressions that are members of this set are about to follow the same path and become members of the set discussed in section 5.2. Although dilly is not recognized as a word that is constructed of or related to dilly-dally or dally, (52) suggests that it is still used that way.

Table

(52)	a.	I can see Aubameyang dillying on any contract extension until […] (NOW 19–09-17 US)
	b.	We had some dillying around and then a hurried throw. (NOW 16–07-10 AU)
	c.	After some dillying, my Spore review for PC Gamer UK went up today. (iWEB)

The expressions in this set display similar behaviour to the expressions in sections 5.1 and 5.2. Their sole meaningful constituents are expressive words. Apart from the imagery of repetition, continuity and alternation, their meanings are similar to the meanings of the constituents, to the point that the items are defined in terms of each other. Wishy-washy and chitchat are characterized as a ‘reduplicated formation on washy adj. (sense 2)’ and a ‘reduplication with vowel variation of chat n.1’ (OED s.v. wishy-washy, chitchat), respectively, while dilly-dally is a ‘reduplication with vowel variation of dally v.’ (OED s.v. dilly-dally). The reduplicative and the single form can also mean the same, as in the case of both chitchat and chat meaning ‘friendly, informal conversation or an instance of this’ (Merriam-Webster s.v. chitchat, chat). There are cases where their meanings are not exactly the same, however. Most definitions of dilly-dally are focussed on aspects of ‘trifling vacillation or indecision’, while dally can be used to characterize other types of behaviour as well; a dilly-dallying spouse need not be a dallying spouse.

For most expressions in this set it holds that the reduplicative and single forms are multicategorial. The constituents can appear in various orders and combinations, even if one order – again the one where the high front vowel is uttered first – seems to be preferred. The constituents can be iterated further mainly as units. Inflectional endings can be realized on the rightmost item, or both items, and the constituents can be joined by overt coordinators. Some examples are given in (53)–(55):

Table

(53)	a.	[…] a retriever that stops short during a retrieve, or dillydallies after picking up the bird, […] (iWEB)
	b.	[…] we appear to have a president who characteristically dillies and dallies, and […] (iWEB)
	c.	Stop your dillying and your dallying, younglings. (iWEB)
	d.	But I dillied and dallied, dallied and I dillied (iWEB)
(54)	a.	So, I cut out the wishy-washing and decided self-pub is the only way to go. (iWEB)
	b.	He s since backtracked a lot, and wishy-washied, and etc. (iWEB)
	c.	He’s both wishy and washy, and […] (iWEB)
	d.	[…] and it gets real wishy and washy and unclear for me […] (NOW 20–07-15, US)
(55)	a.	We chitchat for a while, […] (NOW 10–02-13 NZ)
	b.	We are a small island and it is’ chit chat chit chat’ and […]. (NOW 12–06-13 GB)
	c.	We chitted & chatted, and laughed & laughed […] (NOW 18–0408 CA)

Many, although not all, items in this set are associated with pejorative readings. Such readings mainly result from the constituents describing negative properties. There is, however, a special sub-set where the pejorative readings are a property of the reduplicatives. These are known as schm-reduplicatives (Feinsilver 1961), for example salad schmalad, Toyota schmoyota. The function of the schm-constituent is to signal that a salad is not what one wants, or that Toyota is not the kind of car one thinks highly of. Schm-reduplicatives differ, in terms of meaning and function, from the dilly-dally type items, as well as from most other reduplicative items. They display more compound-like behaviour than any other set we have looked at: there can be exactly two constituents and the schm-constituent is always the second one, no material (and no overt coordinators) can be inserted between the constituents, and material cannot be iterated further. The forms are always associated with deprecative readings, not with imagery of increase in intensity or amount.

5.4. No more shilly-shallying: A look at some remaining cases

We now look at one final set of items that are listed as examples of reduplicative compounds, namely items like willy-nilly and shilly-shally; see e.g. Wheatley (1866), Jespersen (1942: chapter X), Quirk et al. (1972: 1029f.; 1985: 1579f.), Huddleston & Pullum (2002: 1666). These expressions are similar to the harum-scarum and namby-pamby words in section 3.2, in that neither constituent is associated with a meaning in present-day English. Willy-nilly is a variant of the expression will I, nill I, which is based on the word will (OED s.v. willy-nilly). Shilly-shally is constructed of shill I, shall I, which is an ‘altered form of shall I, shall I (not)’ (OED s.v. shilly-shally). Both items are used to signal vacillation and indecision in English, which makes sense in view of how they are formed. As most present-day speakers are probably unaware of this, the items appear very similar to the harum-scarum words in section 3.2. There are, however, important differences between willy-nilly and the harum-scarum words that speak in favour of keeping these items separate. First, although in some of the willy-nilly items, like in the harum-scarum set, the alliteration or rhyme is not associated with a clear function or meaning, this does not apply to all items in this set. (57d) for example implies continuity and a sequence of events. Second, the willy-nilly items are often used as if their constituents were distinct words, and so they display the same behaviour as the items in sections 5.2 and 5.3. Third, in cases where the constituents look identical to existing, etymologically unrelated, words, the reduplicatives may have taken on new meanings that reflect these ‘other’ readings: standardized corpora provide examples of willy nilly, willy and nilly, willies or nillies that are based on willy as a slang word for penis (Merriam-Webster s.v. willy). Examples of how these items are used in present-day English are given in (56)–(57).

Table

(56)	a.	They do it willynilly, it’s their thing. (NOW 16–03-10 GB)
	b.	What the DWP is doing is sanctioning people willy and nilly for not doing arbitrarily-imposed things. (NOW 15–01-08 GB)
	c.	Nothing willy or nilly about that. (iWEB)
	d.	He’s taking care of himself, just in case things get willied or nillied. (Divorced playbook n.d.)
(57)	a.	The shilly-shally continued until the Presidential Convention […] (NOW 22–10-08 NG)
	b.	[…] that she is a shilly-shally person either utterly ignorant or […] (NOW 20–06-02 CA)
	c.	The message is simple: the president shillyshallied on education. (NOW 18–01-25 NG)
	d.	The USA has shillied and shallied because of the Iraq disaster. (iWEB)

The fact that the meaning of the whole is not based on the meaning of either constituent suggests that these expressions, just like harum-scarum, are words in their own right, although in terms of their behaviour the items seem like iterative sequences of words, just like the items discussed in sections 5.2 and 5.3. This may simply be a result of speakers following an existing pattern: willy-nilly, shilly-shally and so on look and ‘feel’ very similar to clink-clank, flip-flop, dilly-dally, and some of the items, like dilly-dally, shilly-shally, are cited as each other’s synonyms (Merriam-Webster, s.v. dillydally, shilly-shally). It is hardly a surprise that speakers use them in the same way as well. A key point that still holds is that the items do not meet the usual criteria for being compound words.

6. Concluding discussion

The aim of the present paper has been to (re-)evaluate the treatment of reduplicative expressions like arty-farty, namby-pamby, hoity-toity, wishy-washy, lardy-dardy and so on as compound words in nearly all previous work on English. We have seen that only a handful of reduplicative expressions meet the criteria for compound words, and even then, they appear to be very atypical compounds. Alongside the compound-like function and interpretation, these reduplicative expressions have another function and interpretation, in which case they display the same behaviour and properties as any other iterative sequences of words in English. For a large majority of reduplicative expressions, this is the main function in present-day English, which means that labelling such expressions compound words is misleading: the expressions are iterative sequences of words, and the iteration has the function of creating emphasis and associations with imagery of intensification, increase, repetition, duration and/or alternation. The sequences are not distinct lexical items that are based on other lexical items. The fact that the number of items in a sequence or their order is relatively free points in the same direction, as does the fact that the individual items can be separated by conjunctions and other connective or linking items, without clear meaning change. This type of behaviour is consistent with the idea that the sequences are created by a syntactic process, which may be the same process that also iterates ‘ordinary’ words in English. For items like pell-mell, harum-scarum, namby-pamby in section 3.2 whose properties and behaviour differ from the properties of most other reduplicative sequences in English we have found, furthermore, that the items are words, but not compound words, as they do not meet any of the criteria for compound words either.

Acknowledgments

I would like to thank the two anonymous reviewers for their insightful and thought-provoking comments on the earlier version of the paper. I would also like to thank Maria Wiktorsson and Carita Paradis for discussions of the topic, and Cian Duffy, Leigh Simpson and Ellen Turner for providing additional native-speaker judgments on some of the examples.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

¹ Some of these terms imply that repetition is a form of dysfluency: involuntary rather than intentional and associated with a function. Involuntary repetition falls outside the scope of the present paper; the same applies to so-called baby-talk, for example weewee and bow-wow.

² Another explanation is the nature of Old English texts: if reduplicatives were part of informal language already at that time, one would not expect them in legal or religious domains, which is where many Old English texts fall.

³ Iterative expressions are not common in standardized corpora: the 14-billion-word Intelligent web-based corpus (iWEB) (Davies 2008c–) contains 50 examples of the spelled form pretty-pretty, 17 examples of prettypretty, and 232 examples of pretty pretty. Wildcard and other open searches add a handful of examples, such as pretty pretties (four occurrences), pretty-pretties (two occurrences), prettyyyy.pretty (two occurrences), and prettypretty.pretty (one occurrence). One reason for the low number is that iterative expressions belong to informal domains, and the sub-corpora that contain data from informal domains are small. As a key aim in the present paper is to illustrate how iterative expressions are used in English – on grounds that most previous work has merely listed them, out of context – corpus examples from big corpora, such as the iWEB, the 18.7-billion-word News on the web (NOW) (Davies 2008d–), or the 1.9-billion-word Global web-based English (GloWbE) (Davies 2008b–), will be provided, whenever possible. In cases where standardized corpora do not contain any relevant examples, web examples will serve as illustrations of usage. Most examples result of lexical searches for items listed in previous work, as we are attempting to either confirm or reject their status as compound words. An anonymous reviewer suggests using acceptability tests as a complement, which indeed would shed light on how the various spelled forms are interpreted. This approach is adopted in a follow-up study that focusses on data like (6b), (10b) or (19)–(20).

⁴ Even pretty is derived using this affix (OED s.v. pretty), although its status as a derived word is less obvious to many speakers.

⁵ The iWEB has one example with three copies: Money, it’s a hit, so don’t give him that goody-goody-goody bullshit. On closer look, this turns out to be part of (discussion of) a song lyric – something that an anonymous reviewer also has noted. Song lyrics, alongside poetry, is a domain where iterative expressions are common, but because they differ from the ‘everyday’ language that we are interested in, they are not used as examples in this paper.

⁶ These sequences should be distinguished from examples like win-win situation where the number of copies is restricted, and the meaning is based on both constituents. In a win-win situation both parties win, and the expression does not imply an increase in intensity, amount or repetitiveness.