The Development of ESP Lexicon Through New Combining Forms

This paper investigates the role of new combining forms in the formation of neologisms which are currently expanding the lexicon of English for Special Purposes (ESP). In the past, only neoclassical combining forms, such as initial bio-or final -logy (in biology ), were productively used in ESP. Nowadays, specialized combining forms also include abbreviated forms of existing words (e.g., cyber-from cybernetic in cyber-attack ), as well as secreted (i.e. reinterpreted) forms (e.g., -bot from robot denoting ‘a type of automated program or software’ in knowbot ). The paper explores a set of combining forms attested since the second half of last century in the online version of the Oxford English Dictionary (OED) with the aim to demonstrate how specialized sectors, such as science or information technology, are being enriched by series of combining-form combinations. The paper conducts quantitative analyses in the Corpus of Contemporary American English (COCA) and the News on the Web Corpus (NOW) to substantiate the frequency and stability of specialized combining forms and their profitability in the formation of both novel and nonce words.


Introduction
Because of new scientific discoveries, medical advancements, and technological inventions, the lexicon of English is in continuous expansion (Arndt-Lappe et al., 2018).Many new words are recognized and repeatedly used by the speech community, therefore becoming part of English vocabulary as accepted neologisms.Others are coined for specific occasions and do not always become institutionalized or lexicalized (Brinton & Closs Traugott, 2005), yet their creation suggests how language may be enriched with new temporary vocabulary, namely nonce terms, connected to a historical, socio-political or economic period.The lexicon of English for Special Purposes (ESP), however, is rarely ephemeral, since the emergence of new specialized terms is generally motivated by lexical gaps which need to be filled.
The present paper analyses 30 combining forms retrieved from the Oxford English Dictionary (OED) and contributing to the lexical innovation of ESP.Combining forms (henceforth, CFs) are traditionally defined as neoclassical elements of Greek or Latin origin, such as bio-(Gr.βίο, Lat.bio 'life') or -logy (Gr.λογία, Lat.logia 'science'), which are added to other elements, either initially, as in biodata 'one's curriculum vitae', or finally, as in musicology 'the branch of knowledge that deals with music'.However, as recently demonstrated by Mattiello (2022), CFs comprise not only neoclassical bound elements, but also abbreviated forms, such as e-(from electronic) in e-book, e-journal, and secreted ones, such as -nomics (from economics, denoting 'the economic policies of a President or head of state') in Clintonomics, Obamanomics (cf."classical" vs. "modern combining forms" in Prćić, 2005Prćić, , 2008;;Amiot & Dugas, 2021).
While in the literature neoclassical CFs are generally regarded as cases of compounding (Bauer, 2017, p. 150), abbreviated and secreted CFs have been attached different labels, including "affixoids" or "pseudo-affixes", which rather stress their derivational nature.Scholars also tend to confuse or merge CFs with the notion of "splinters" used in the formation of blends (Bauer, Beliaeva, & Tarasova, 2019, p. 62), thus confirming their "fuzzy" nature and difficult categorization.In a recent study (Mattiello, 2022), the three categories of neoclassical, abbreviated, and secreted CFs have been considered as part of transitional morphology, i.e. intermediate between two subcomponents of word-formation: i.e., derivation and compounding.Like affixes, CFs are bound morphemes added to bases, yet like compound constituents, they have a high lexical density and can be stressed (e.g., e-has the same meaning as the adjective electronic from which it is clipped and is stressed in é-mail).
The dividing line between one category and the other is often not so well defined, nor are CFs classified in a univocal way in the linguistic literature.Section 2 tries to clear up the terminological confusion about CFs and related concepts.Section 3 explains the methodology followed for the selection and analysis of the data.Section 4 offers a qualitative scrutiny of the selected CFs and a quantitative analysis of the combining-form combinations that they produce, discriminating between highly frequent neological formations and occasionalistic nonce words.Finally, section 5 summarizes the research conducted and draws up some conclusions.Warren (1990, p. 112) was the first to stress that "combining forms are morphemes of a rather special kind", which represent a quite heterogeneous group of elements, differing in their origin as (1) allomorphs of model words (e.g., astro-in astrodome), (2) truncated forms of model words (e.g., cyber-in cyberphobia), or (3) parts of model words which happen to coincide with existing words (e.g., -gate, from Watergate, denoting 'an actual or alleged scandal' in sexgate).The same CF can even differ in the way it is interpreted depending on the word to which it is added: e.g., in cheeseburger 'hamburger topped with cheese', -burger is an abbreviated form, whereas in fishburger 'fried patty made of fish served in a bread-bun', it is secreted, in that a fishburger does not contain any beef meat.Moreover, CFs are divided into initial and final (e-vs.-nomics) depending on their position either to the left or to the right of a base.With regard to position, there are very few exceptions (i.e.morph-/-morph and phil-/-phile), which can occur both in initial (e.g., morph-ology, phil-ology) and final positions (e.g., anthropo-morph, anglo-phile) (Bauer, 1983;Plag, 2003).Initial CFs are close to prefixes while final CFs are close to suffixes.This heterogeneity manifests itself in terminological confusion (e.g."affixoids", "pseudo-affixes", "semi-affixes", "affix-like formatives", "folkmorphs", "splinters", along with "combining forms") and descriptive vagueness.Bauer already in the 1980s claimed that CFs belong to a type of word-formation which "has received very scant attention in the literature on morphology" (Bauer, 1983, p. 213).Scholars indeed tend to relegate this topic to the category of neoclassical CFs and only include discussions on neoclassical compounds in their description of English word-formation, disregarding completely the modern categories.More recently, Kastovsky (2009, p. 12) has even argued that the notion of CF is not necessary at all, as "[t]he categories of 'word ', 'stem', 'affix', 'affixoid', 'clipping' and 'blending' necessary in word-formation for independent reasons are sufficient to deal with the formations in question".

Combining Forms: Definitions and Terminological Distinctions
Following recent studies on the topic (e.g., Amiot & Dugas, 2021), Mattiello (2022) has dealt with this phenomenon showing the importance of a finer-grained classification and an in-depth description of CFs.From her analysis, CFs consist of a miscellaneous set of morphological elements which share features such as semantic and phonological weight with compound constituents, but also features such as boundedness and fixed position (either before or after a base) with affixes.For these reasons, Mattiello (2022) accommodates the different categories of CFs under the same umbrella term of "transitional morphology", straddling the demarcation line between two subcomponents of word-formation, i.e. derivation and compounding.
In line with Mattiello (2022), CFs are here defined as initial or final bound morphemes which are either allomorphic variants of classical Latin or Greek words (e.g., bio-above), or shortenings of (native or non-native) English words (e.g., econo-from economic in econo-politics), often with the intervention of a secretion process (e.g., -aholic 'person addicted to' in work-aholic 'person addicted to work').These are respectively called "neoclassical", "abbreviated", and "secreted CFs", and the complex words obtained from them are called "combining-form combinations".
A notion which is often confused with CFs is that of affixoids.The term "affixoid" is used within Construction Morphology by Booij and Hüning (2014), who note how the element free has developed the more general meaning 'without what is denoted by the base word' when it is used as the right constituent of compounds, as in sugar-free and fat-free.Moreover, unlike the suffix -less having the same denotative meaning in sugar-less and fat-less, the affixoid -free also presupposes a slightly negative evaluation (e.g., that the presence of sugar is not good for one's health).Therefore, affixoids exist alongside formally identical and usually free 'parent' morphs; yet, they acquire a more generalized meaning (Stevens, 2005, p. 73) or undergo a "desemanticization" process (Amiot & Dugas, 2021), as -free 'clear of something which or someone who is regarded as problematic' in trouble-free, pollution-free, risk-free, etc.Thus, like CFs, affixoids are "productive" (Bauer, 2001), i.e., are part of schematic constructions with high type frequency (e.g., [N-free]), and are bound to their bases (e.g., in sugar-free, the suffixoid and the base cannot be split up as *sugar very free).However, unlike CFs, affixoids correspond to free-standing words, whereas in the case of CFs, correspondence to a free morpheme is accidental or synchronically irrelevant (cf. the CF -gate 'political scandal' vs. the noun gate).
Still another crucial distinction is between CFs and blend splinters.Splinters are word parts that merge with one another or with full words in a blend (e.g., in jeggings ← jeans + leggings there are two splinters which graphically overlap, while in Brexit ← British + exit only the first element is shortened).Hence, the notion of splinter is very close to that of CF, differing from it only in their diverse degrees of productivity (Bauer, 2001).We may even envisage a diachronic evolution from blend splinter to abbreviated or secreted CF.Many splinters are indeed transitional from their status as blend constituents (i.e.word parts) to a more productive affix-like status as CFs (Correia Saavedra, 2013, 2016).This evolution is manifest in -aholic, which was originally attested in blends (e.g., work-aholic [1947], choco-holic [1961]), then became a productive CF included in the OED with the meaning 'a person addicted to - ' (e.g., carbo-holic [1973], shop-aholic [1977], news-aholic [1979], spend-aholic [1982]), with a higher degree of frequency, productivity, and stability.
In this paper, the focus is on the CFs which have become productive and stable in ESP, because of their reiterated use in the formation of novel words by experts in specialized areas.

Dataset and Methodology
The methodology adopted for the selection of the CFs examined in this study combined an advanced search in the OED, using the designated tool, with a manual cleaning of irrelevant examples.The advanced search was conducted in November 2023.First, I selected the tab 'entries' of the OED and entered the search text 'combining form'.This search returned a list of 2,297 matching entries, which were then put in chronological order, by arranging results by 'date (newest first)'.Since my focus was primarily on recent forms, I added a temporal parameter to my search, by selecting the range 1950-today, and concentrated on this dataset.The 77 results obtained were the starting point for my analysis.To this list robo-was added manually because, even if its earliest attestation is before 1950, in 1988 it became a secreted form referring to something or someone 'resembling a robot, esp. in being resilient, emotionless, or futuristic'.Another CF that deserved inclusion, but did not match my advanced search because it is not dated in the OED, is -nomics.I decided to incorporate it because the earliest formation displaying this CF is dated 1969.
The final 79 results include all three categories -namely, neoclassical, abbreviated, and secreted forms -yet with some elements having evolved in time from abbreviated to secreted.This dataset was then manually reduced to 30 -18 initial and 12 final -specialized CFs for the purposes of my analysis.Among the selected CFs, some are apparently similar in form, but actually diverge in meaning: e.g., e-is abbreviated from either electric (e.g., in e-car) or electronic (e.g., in e-journal).Hence it counts as two separate CFs (e-1 vs. e-2 ) used in different specialized fields.The specialized fields of the selected CFs mainly involve digital and information technology, electronics, economics, and less relevantly chemistry, medicine, physiology, palΘontology, and linguistics.
In the corpus-based analysis, the three categories of neoclassical, abbreviated, and secreted CFs will be treated separately, also distinguishing between initial and final for each type, in order to verify their frequency and profitability in terms of new words obtained through this word-formation mechanism.Both type and token frequencies of occurrence in the Corpus of Contemporary American English (COCA) (Note 1) and in the News on the Web Corpus (NOW) (Note 2) will be shown, as well as their meanings and contexts of use.By observing quantitative data, a distinction between stable neologisms and nonce formations or hapax legomena (i.e.once-only attestations) will be also made.By exploring qualitative results, an evolution towards abbreviated and secreted CFs will be observed in ESP, as already remarked in general English (Mattiello, 2022).

Specialized Combining Forms in ESP: Stable and Unstable Formations
Specialized CFs are certainly expanding the lexicon of ESP with new words.The aim of this section is to establish the productivity of these morphological elements used in different areas and with diverse degrees of stability.Some of them are more stable and produce large sets of new words which deserve inclusion in lexicographic works, others, being more specific and technical, have a limited applicability, but still contribute to the lexical innovation of ESP.

Neoclassical Combining Forms
Neoclassical CFs mainly pertain to the areas of chemistry, biochemistry, and medicine, with rare exceptions.
A few of the new words are hapaxes (e.g., ichnofamily and ichnographic(ally)), yet most of them are stable neologisms (e.g., This confirms the capacity of the body fossil record and ichnology to complement each other.NOW, 17/03/2023).
Another specialized CF coming from Greek λεξικός 'lexicon' is initial lexico-, which has been attested from 1953 in modern linguistic terms denoting 'lexical and -'.It often combines with adjectives (e.g., lexico-behavioural, lexico-grammatical), but especially occurs before nouns (e.g., lexico-dynamics) or final CFs (e.g., lexicography).Nouns such as lexicogrammar, lexicology, lexico-statistics occur more frequently than the derivatives lexico-grammatical, lexicologist, lexicostatistical, lexico-statistician, with only one occurrence in either COCA or NOW.Nevertheless, many of these words (e.g., lexicographer, lexicography) are amply attested in dictionaries and corpora (e.g., Davis studied at Rhodes University and Oxford before working in lexicography at the Oxford English Dictionary.NOW, 07/01/12).
In chemistry, seco-(recorded since 1951) is a formative element used in naming derivatives, especially of steroids, in which fission of a ring has occurred, as in seco-dicarboxylic and seco-steroid.The origin of the CF is from Latin secāre 'to cut', followed by an -o-connective (Note 3).The use of this initial CF is very scanty, though, as the only word attested in corpora is seco-steroid (e.g., The body metabolizes it into a seco-steroid hormone which, like all steroids, regulates genetic transcription, a basic and profound mechanism of action unlike any other vitamin.NOW, 14/06/18).
Final CFs are more frequent in corpora than initial CFs in terms of both types and tokens.In chemistry and biochemistry, the final CF -mer (from ancient Gr. -μερής 'part, segment', first attested in 1975 in the OED) forms nouns denoting particular kinds of polymer (as in dimer, oligomer) or isomer (as in epimer, tautomer).It can also be suffixed to numerals (in full form or more usually as digits) to form nouns denoting polymers consisting of a given number of units (as 9-mer, 16-mer).Combinations with -mer mainly include stable words such as copolymer, monomer, polymer, tetramer, trimer (e.g., This implied that the complex consisted of three identical subunits, arranged as a symmetrical trimer.COCA, 01/08/18), with rare nonce words or hapaxes such as 25-mer and 36-mer.
Another frequent final CF is -ogen, first attested in 1961 in the OED and coming from an -o-connective and the Greek root γίγνεσθαι 'be born, become'.In biochemistry, it forms the names of an inactive precursor of a compound to whose name it is appended, as in hypertensinogen, kininogen, etc. English has inherited from French the names hydrogène and oxygène, which were adopted into English with the ending -gene, afterwards altered to -gen.On the analogy of these words, a considerable number of new terms have been added to the English vocabulary of chemistry, in which the ending -(o)gen expresses the sense 'that which produces'.They are usually names of chemical substances, as in amidogen, cyanogen, nitrogen, etc., or of classes of substances, as in halogen.The most frequent words in both corpora are hydrogen, estrogen (e.g., The basic components of HRT are estrogen and progestin, which is a form of progesterone.NOW, 10/11/23) and nitrogen, but novel nonce words such as angiogen, astrogen, carbogen, entactogen, vitrogen also demonstrate the profitability of -ogen in the coinage of new vocabulary.
The CF -penia, first attested in 1971 in the OED, originates from ancient Greek πενία 'poverty' and is primarily used in medicine.From the end of the nineteenth century, it has been found in a small number of English formations and adaptations of foreign words, apparently first in leucopenia, and later in granulocytopenia and neutropenia.It chiefly combines with first elements of Greek origin, as in leukopenia/leucopenia, osteopenia, sarcopenia (e.g., Recent estimates suggest that sarcopenia affects 10% to 16% of the elderly population worldwide.NOW, 13/10/23), and thrombocytopenia, which are stable formations, unlike the hapaxes cytoglucopenia, erythroblastopenia, plancytopenia, reticulocytopenia, and vasopenia attested only once in the corpora.
The CF -valent, from Latin valentem (present participle of valēre 'to be worth'), has been attested in the OED from 1977.It is a formative element occurring in a few words of general currency and in various scientific contexts, used with prefixes denoting number.For instance, in chemistry and immunology, it forms adjectives denoting 'having a valency of the specified number ' (e.g., trivalent, quadrivalent, tetravalent).It is also used with Arabic numbers prefixed (e.g., 14-valent).While ambivalent (with the prefix ambi-), covalent, multivalent, and quadrivalent are amply attested in the corpora (e.g., A quadrivalent vaccine works by stimulating an immune response against four different antigens, such as four different viruses or other microorganisms.NOW, 09/02/23), attestations of hexacovalent, hypovalent, omnivalent, and plurivalent are rare occurrences.
Table 1 summarizes the quantitative data resulting from the COCA and NOW corpora showing the productivity of specialized neoclassical CFs in terms of type frequency, token frequency, and hapax legomena.The high frequency of CFs such as initial lexico-and synapto-and final -ogen and -valent shows the "realised productivity" or "extent of use" (Baayen, 1993) of such morphological elements and their profitability (Corbin, 1987;Bauer, 2001, p. 49) for the expansion of ESP lexicon.The fact that these are also used to coin nonce terms attested only once in corpora (hapaxes) signals their "expanding productivity", measuring the rate at which these morphological categories are expanding and attracting new members.Some neoclassical CFs (e.g., ichno-and seco-) are quantitively less productive, but their lower frequency is correlated to a sector-related use and specificity.

Abbreviated Combining Forms
Abbreviated CFs do not have a classical origin, but are obtained by shortening existing words and combining them with bases.Many of them are used to coin new words in the areas of digital technology, electronics, and economics.
Among initial CFs, one of the most frequent is cyber-.Shortened from cybernetic, it was originally (from 1961) used to create words, mainly nouns relating to computers, information technology, and virtual reality, or denoting futuristic concepts.Later, cyber-was specifically used to form terms relating to the Internet, probably influenced by analogy to cyberspace.Formations include established terms, such as cyberart, cyberattack, cybercommunity, cybercrime, cyberfriend, cyberlover, cyber-romance, cybersecurity, cyberspeak, cybersphere, cyberterrorism, and cyberworld, as well as temporary nonce words (e.g., cyber-family, cyber-debate, cyber-piracy, cyber-protest, cyber-secure) whose number is daily increasing.As a separate entry, the OED offers another initial CF e-2 , more recent [1988] and this time abbreviated from electronic.Found in numerous recognized formations from the late 1980s, as in e-book and e-fit, early formations are often preceded by corresponding compounds in electronic whose spelling is as two separate words (e.g., electronic book [1978] vs. e-book [1988], electronic cash [1967] vs. e-cash [1994]).
The CF e-2 is indeed prefixed to nouns to indicate involvement in electronic media and telecommunications (esp.concerning the Internet), usually to distinguish objects or actions from their non-electronic counterparts.In the late 1990s and early 2000s, formations starting in e-2 became very numerous and conspicuous in their frequency of use.The three specific semantic areas that these formations cover are: (1) terms relating to the publication or exchange of information in an electronic format (e.g., e-edition, e-journal, e-publication, e-reader, e-text, e-zine); (2) terms relating to computers and the Internet, i.The initial CF petro-1 , shortened from petroleum, first occurred in the mid-twentieth century in petrochemical [1942] and petrochemistry [1942], although petromax (from G. Petromax) [1929] 'a proprietary name for a type of vaporised paraffin pressure lamp' and Petro-Forge [1969] 'a kind of forging machine powered by a petrol engine' also belong here.As an abbreviated form, it obtains terms relating to petroleum, especially with reference to the political and economic power of oil-producing countries, as in petropolitics (e.g., Petropolitics come with financial risks as well as upsides.NOW, 04/01/23), petro-power, petro-resources, and petro-wealth.Formations abound in corpora, including, among others, both petrochemical, petrography, petrologist, petrology, petromania, petropolitics, petro-state, and a conspicuous number of nonce words (e.g., petro-domination, petro-fueled, petro-military, petro-nation, petro-oil, petro-plastic, petro-supplier, etc.).
The CF porta-, shortened from the adjective portable, apparently appeared earliest in portapak [1951] 'a portable system comprising a video camera and recording equipment'.It is used to form the (often proprietary) names of various movable or portable versions of manufactured objects, devices, structures, etc., as specified by the second element: e.g., portacam, portacrib, Porta-John (from slang john 'a toilet, lavatory'), porta-kit, porta-office, Porta-Phone, porta-printer, porta-screen, etc.In corpora, we also find portaball, portacabin (e.g.aerobot, autobot, chatbot, guardbot, medbot, microbot, minibot, newsbot, repairbot, sexbot, smartbot, teacherbot, twitterbot, waterbot, whereas agrobot, brainbot, girlbot, megabot, and securitybot are some instances of nonce formations. Attested in the OED from 1965, the abbreviated CF -lect (from dialect) is a terminal element used to designate a regional or social variety within a language, as in idiolect [1948] 'the linguistic system of one person, differing in some details from that of all other speakers of the same dialect or language'.It is also used in forming a number of technical terms in linguistics, as in acrolect, basilect, isolect, sociolect, etc.Hence, it has been converted to a noun and used to refer to 'a social variety of a language or dialect'.The most common formation in both corpora is idiolect (e.g., Attackers can mimic the distinct idiolect of the target.NOW, 18/11/23), followed by Afrolect, basilect, ethnolect, interlect, multilect, sociolect, while hyperlect and topolect are hapaxes.
The final CF -olol (from propranolol) is found in formations from the late 1960s onwards.It is typical of pharmacology, where it is used to form the names of 'beta-adrenoceptor blocking drugs derived from propranolol or having a similar molecular structure' (e.g., acebutolol 'a beta blocker which is partially selective for cardiac beta receptors, used in the treatment of hypertension, angina, and arrhythmia').Specialized formations attested in corpora include atenolol, bisabolol, bisoprolol, metabolol, metoprolol, nadolol (e.g., We aimed to test the effectiveness of isosorbide mononitrate as an adjunct to the b-blocker nadolol in the prophylaxis of first variceal bleeding in these patients.COCA, 21/12/1996), nebivolol, stanozolol, timolol, while nandolol, penbutolol, pindolol, practolol, and talinolol are once-only occurrences.
In particle physics, -onium represents the abbreviation of positronium.Since 1987, it has been used to form the names of bound states of a particle and its antiparticle, as in charmonium, nucleonium, toponium.This is not to be overlapped with the homonymous antecedent form, occurring in chemistry from 1858 in carbonium, hydrazonium, nitronium, phosphonium, which instead comes from ammonium and forms the names of 'complex cations that contain a more or less electronegative central atom'.The only stable word attested in corpora is charmonium (e.g., And, given that the specific quark is called "charm," its existence opens vast possibilities for puns -and that's without even getting into the fact that the technical term for the full family of particles containing these quarks is "charmonium".NOW, 14/07/20), while quarkonium is a once-only formation attested in NOW.
The CF -ylidene, from ethylidene, earliest attested in 1971, is used in chemistry as an adaptation of a French chemical name.It is used to form the names of divalent radicals in which both valencies derive from the same atom, as in benzylidene.It replaces the -idine suffix when the name of the parent compound does not end in -yl (cf.alkylidene, propylidene, etc.).Corpora attest stable words such as 2-chlorobenzylidene, 4-methylbenzylidene (e.g., An order came into force on Wednesday banning lotions containing oxybenzone, octinoxate, 4-methylbenzylidene camphor or butylparaben from Thailand's marine national parks.NOW, 04/08/21), benzylidene, cyclopentylidene, nitrobenzylidene, o-chlorobenzylidene, retinylidene, but 2-xhlorobenzylidene is attested only once in NOW.
Table 2 summarizes the quantitative data resulting from the COCA and NOW corpora showing the productivity of specialized abbreviated CFs in terms of types, tokens, and hapaxes.As Table 2 shows, specialized abbreviated CFs are highly frequent in corpora, particularly initial cyber-and e-2 , productively forming hundreds of types, but also initial petro-1 and econo-, and final -bot 1 , used to form both neologisms and nonce words.The high number of tokens occurring in corpora and the numerous hapaxes obtained through these morphological elements confirm the stability of the CFs and their acceptability for formations and vocabulary expansion.

Secreted Combining Forms
Secreted CFs involve both abbreviation and reinterpretation.Formally, they are obtained from the back-or fore-clipping of a word.Semantically, they involve either a generalization or a specification process, which contributes to conferring a certain level of abstraction on the CF.Many of them are used to coin new words in the areas of digital technology and, less frequently, economics.
An initial CF which is not only abbreviated (see its concrete use for digital devices in § 4.2) but also secreted is digi-2 .Attested from 1986 as a secreted CF, it is used to create 'nouns denoting a product, process, person, etc., relating to or characterized by the use of digital technology', such as digi-age, digi-art, and digi-novel.On some occasions, it is also prefixed to the names of styles of popular music (as in digi-funk, digirock, etc.) to form 'nouns denoting music incorporating sounds generated or modified digitally'.Hence, unlike digi-  , digiconomics, digiconomy, digimarketing, digi-paint, digiphile, digishop, digitour, digizine, while digidance, digidata, digi-friendly, digimusician, digirama, and digitask are nonce words.
The formation of the secreted CF m-, from mobile, is clearly analogical: its similarity with e-2 (see § 4.2) is marked formally -both retain only the initial letter of a word, being rather borderline with initialisms -and semantically.Recorded earliest in m-commerce [1997], m-forms terms relating to 'commercial activity conducted through mobile electronic media and devices, especially mobile phones', as in m-banking, m-payment, m-ticket, etc.
Since 2001 it has also been used to create terms (occasionally temporary words and ad hoc formations) relating to 'social and cultural activity or phenomena conducted through or using mobile electronic media and devices', as in m-government, m-health, and m-voting, but earliest in m-learning.The analogical nature of these formations is confirmed by the existence of e-2 parallel formations (cf.e-commerce, e-learning) (e.g., Share of m-commerce into e-commerce is projected to reach 80 percent by 2024.NOW, 29/11/23).Other recognized formations are attested in the COCA and NOW corpora: e.g., m-cash, m-enterprise, m-finance, m-government, m-retailing, m-service, m-system, m-ticket, while hapaxes in COCA include m-banking, m-finance, m-healthcare, m-mailbox, and m-money.Besides being an abbreviated initial CF shortened from petroleum (e.g., petrochemistry, see § 4.2), petro-2 is also secreted when it forms terms designating 'revenue, esp.foreign exchange, that derives from petroleum exports', as in petrodollar [1973] 'a notional monetary unit earned by a country from the export of petroleum', petrocurrency [1974] 'the currency of a petroleum-exporting country', petrobillion, petro-naira, and petropound.Table 3 indicates that secreted CFs are less frequent than neoclassical and abbreviated CFs, in that they represent a rather recent phenomenon, still ongoing and sometimes developing from blends or abbreviated forms.In most cases, indeed, shortening precedes the semantic reinterpretation process, as with digi-2 , petro-2 , and -bot 2 , all corresponding to antecedent abbreviated forms (see § 4.2).The scarce number of hapaxes also goes in the same direction, confirming that secreted CFs are contributing to the development of ESP less conspicuously than neoclassical and abbreviated forms.Nevertheless, some less specific CFs, like initial robo-, are producing both types and hapaxes, showing an increasing productivity and going towards stability and recognition.

Conclusions
The present study has investigated the role of combining forms in the development of ESP lexicon.Although all the CFs examined here are recorded in the OED, their productivity is increasing and the stability of the combining-form combinations resulting from them has been confirmed by corpus linguistic analysis.
Corpus-based results have demonstrated that, besides neoclassical CFs already attested in ESP since the nineteenth century, abbreviated and secreted CFs are also currently contributing to the lexical expansion of specialized English.
Some of the CFs analyzed have formed series of new words, especially stable and recognized neologisms, but also nonce words or hapax legomena which are attested only once in corpora, thus showing the potential profitability of these morphological elements in the coinage of still novel formations.
CFs are transitional phenomena often originated from a blending process, or from a shortening process followed by semantic reinterpretation.Secreted CFs, which involve semantic specialization or generalization, represent the last step of the morphological evolution from blend splinters to productive morphological elements that are very close to affixes.Hence, they are less frequently found in corpora, but their expanding productivity should be monitored in order to show how relevant they are to the vocabulary expansion of ESP.
The cases of net-(from Internet) and -flation (from inflation) belong here, in that they are splinters originated from blends (e.g., netiquette ← Internet + etiquette; stagflation ← stagnant + inflation), but they are becoming productive in the formation of novel words, such as slumpflation and netizen or Netscape (Note 5).Thus, the two splinters are potential CFs to be recorded in dictionaries not as part of blends, but as separate entries (i.e. initial CF net-'related to Internet' and final CF -flation 'increasing inflation accompanying the state of the economy'), in spite of the scarce number of novel formations found in corpora.
Hapaxes in corpora, indeed, are often attested only once because of their sector-relatedness and specificity, but tend to be recognized very soon by the community of experts who are familiar with the lexicon of ESP and to be recurrently re-used in professional contexts as a confirmation of their lexicalization and stability.
Note 5.The OED records the following blends: netiquette [1982] 'an informal code of practice regulating the behaviour of Internet users when using email, bulletin boards, chat rooms, newsgroups, etc. ', netizen [1984] 'a person who uses the Internet, esp.habitually', Netscape [1988] 'a proprietary name for: a browser used to access and display documents on the World Wide Web and other computer networks'; stagflation [1965] 'a state of the economy in which stagnant demand is accompanied by severe inflation', slumpflation [1974] 'a state of economic depression in which decreasing output and employment in industry are accompanied by increasing inflation'.

Table 1 .
Frequencies of occurrence of neoclassical CFs in the COCA and NOW corpora The following example illustrates cybercrime in context: In this age of rapid technological advancement and widespread internet usage, cybercrime has emerged as a grave menace to our society, economy and personal security (NOW, 23/11/23).Shortened from the adjective digital, digi-1 is an initial CF attested from 1960 to form nouns denoting 'a digital device, system, etc.', as in digibox, digicam.Later its meaning has been reinterpreted and, as a result, the initial CF has become secreted (cf.§ 4.3).Besides digibox and digicam (e.g., Film high DR roll-off in the highlights is sorely missed with digicams.NOW, 21/10/19), corpora attest the use of neologisms such as digicode, digicorder, digipack, digiPad, digitech, as well as nonce words occurring only once(e.g., digicell, digi-drum, digiPod,  digi-spoon, and digitape).announced, the initial CFs e-1 and e-2 are kept distinct because of their different origin and use.The initial CF e-1 is earliest attested in 1969 with the meaning electric, from which it is abbreviated.With this sense, it forms nouns denoting vehicles powered by electricity instead of or as well as more traditional means, as in e-car, e-bicycle (or e-bike), and e-scooter.In corpora, only stable nouns occur, such as e-boat, e-bus, e-cycle, e-ferry, e-jeep, e-power, e-rail, e-street, e-train, and e-van, although e-car and e-scooter recur more frequently (e.g., The launching of the e-scooter follows the commencement of the construction of Electrum's electric scooter factory in June 2023.NOW, 23/11/23).

,
Portacabin schools were started by state government in 2012 to cater to tribals.NOW, 19/01/23),porta-cell,  porta-chef, portal-gun, porta-potty, and hapaxes such as portabook, porta-box, portacomstyle, porta-filter,  portaphone, portapool, and portaportal.The CF syn-, abbreviating synthetic, has been used from 1971 to form words denoting synthetic products, as in syncrude 'a synthetic product made from coal in imitation of crude oil', synfuel, syngas, synjet, synoil, and  synroc.As such, it differs from the prefix syn-, a Latinized form of Greek συν-'together' (e.g., synergy, synthesis), and generally carries stress.In corpora we mainly find stable formations, such as syncrude, syndiesel, syn-free, syngas (e.g., Electrochemical splitting of CO 2 can create useful chemicals, such as syngas, Formic acid, Oxalic acid, and more.NOW, 30/10/23), synlawn, syn-turf, and some rare nonce formations (e.g., synsilk, synskin).The CF -bot 1 (from robot) is found in a number of formations from about 1966 onwards.The new words were originally blends involving robot (e.g., mobot from mobile and robot), but later -bot 1 was used to form nouns denoting 'a type of robot or automated device' or, in extended (metaphorical) use, to refer to 'a person regarded as an automaton', as in fembot[1976]'a robot resembling a woman in appearance' and nanobot [1989]'a nanorobot' (e.g., The helical nanobot is made of silicon dioxide coated with iron.NOW, 16/05/22) (cf.its specialized meaning in information technology, § 4.3).Other stable formations attested in corpora include

Table 2 .
Frequencies of occurrence of abbreviated CFs in the COCA and NOW corpora 1 , digi-2 does not refer to devices, but to more abstract entities such as art, age, or music, as in digisphere (e.g., You can just launch and then let it disappear into the digisphere.NOW, 02/06/19).Other stable formations attested in corpora include digibook The secretion process activates semantic specification in this case.The most common word in corpora is petrodollar (e.g., Since the menace of armed robbery went full throttle in the immediate petrodollar Nigeria of the early 1970s, it has grown further into becoming a social pandemic today.NOW, 07/05/23), yet other formations are attested: e.g., petrobank, petro-billionaire, petro-currency, petroeuro, petromoney, petroyuan, and the nonce terms petro-bourgeoisie, petro-charged, petro-shilling, petro-wealthy, and petro-welfare.The initial CF robo-, clipped from robot, has been attested from 1988 in a number of formations with the sense 'a/an -resembling a robot, especially in being resilient, emotionless, or futuristic'.Examples include both recognized nouns, such as robocop 'a robotic or bionic law enforcement officer' (used in science fiction from 1957) and robocall[1998]'an automated telephone call which delivers a recorded message' (e.g., But on Sunday Khatari cut a robocall to voters urging them to vote for Kagan.NOW, 06/11/23), and nonce words such as robo-boxer, robo-candidate, and robodisco.In corpora, we also find robo-adviser/robo-advisor, robocaller,  robocar, robocat, robodoc, robodog, roboform, robohelp, robonaut, robo-signer, robosigning, etc.Nonce words are surprisingly frequent, a few instances being robobike, robobox, robobug, robobuilder, robodoctor, robohunter, robokiller, robomom, and robostaff.Shortened from the same base robot, the final CF -bot 2 is instead used in computing and information technology to form nouns denoting 'a type of automated program or (Internet) software, especially one which searches out information' (cf.also abbreviated -bot 1 in § 4.2).More established combinations of this nature are infobot[1986]'any of various automated systems for providing or obtaining information', knowbot[1988]'a program designed to search through large numbers of databases and retrieve information', and cancelbot[1993]'a program that searches for and deletes specified postings from Internet newsgroups'.Corpora also attestApplebot, biobot,  chatterbot, eventbot, filebot, Googlebot, junkbot, shopbot, spambot, talkbot, Twitterbot, etc.The final CF -verse (from universe) has been used from 1981 to form nouns denoting 'the sphere or realm of what is specified or indicated by the first element'.Before becoming a recognized CF, the blends nulliverse [1847] 'a world devoid of any unifying principle or plan' and multiverse [1895] 'the universe considered as lacking order or a single ruling and guiding power' occurred.Nowadays, -verse is frequently used in the context of computing and electronic communications, as in blogiverse, digiverse, and Twitterverse (e.g., The Twitterverse further said that the airline just followed the rules.NOW, 12/08/23).Other formations are attested in corpora:Appleverse, megaverse, metaverse, negaverse, pluriverse, televerse, webverse, etc.However, from  1993, it has also obtained nouns denoting 'the fictional world associated with a specified character, television series, author, etc.', as in Potterverse (from J. K. Rowling's Harry Potter series), Honorverse (from David Weber's Honor Harrington series), Buffyverse (from the black comedy fantasy film and later television series Buffy the Vampire Slayer), Whoniverse (from the television series Doctor Who), and the nonce terms Bronteverse, Garfieldiverse, Marvelverse.Therefore, the CF acquires more specific meaning in the secretion process.
Some of the less common or more ephemeral forms are searchbot, spybot (e.g., Finish them, and grab the Spybot before sliding down to the bazaar.NOW, 10/06/21), and warbot.A semantic specification process has intervened in the formation of this CF.The final CF -nomics is found in formations from the first half of the nineteenth century in the sense denoting the science or study of a subject specified by the first element, as pyronomics, phoronomics, etc.However, from the second half of the twentieth century, it has formed nouns denoting (often semi-humorous) fields of economics, from which it is shortened, as specified by the first element.It is therefore reinterpreted, via a specialization process, as 'the economic policies of -', the second element often being the name of a president of the United States, as in Nixonomics[1969](from Richard Nixon), Reaganomics[1970](from Ronald Reagan), and Clintonomics[1992](from Bill Clinton), or another famous leader or political authority, as in Rogernomics[1985](from the name of Roger Owen Douglas, New Zealand Minister of Finance).Other attested formations include Abenomics,Berlusconomics, Bushonomics, Carternomics, Gandhinomics,  Obamanomics, Osbornomics, Popenomics, Putinomics, Thatchernomics, Trumponomics (Note 4), and the nonce terms corruptonomics, Dixonomics, dole-nomics, Dollarnomics, Gorbanomics, Wikinomics, womanomics, some of which extend the base to a common noun (i.e.corruption, dollar, woman).A very recent formation occurring 2,186 times in NOW is Bidenomics (e.g., U.S. President Joe Biden touts the administration's "Bidenomics" agenda during an October speech... from the White House.NOW, 22/11/23).

Table 3
summarizes the quantitative data resulting from the COCA and NOW corpora showing the productivity of specialized secreted CFs.The absence of hapax legomena in NOW is generally related to the fact that the automatic search is limited to 1,000 results per CF.

Table 3 .
Frequencies of occurrence of secreted CFs in the COCA and NOW corpora