English suffixes

Stress-assignment properties, productivity, selection and combinatorial processes

by Ives Trevian (Author)
©2015 Monographs 471 Pages
Series: Linguistic Insights, Volume 202


English morphophonology has aroused considerable interest in the wake of Chomsky and Halle’s ground-breaking The Sound Pattern of English (1968). Various theoretical models have subsequently emerged, seeking to account for the stress-placement and combinatorial properties of affixes. However, despite the abundance and versatility of research in this field, many questions have remained unanswered and theoretical frameworks have often led their proponents to erroneous assumptions or flawed systems. Drawing upon a 140,000-word corpus culled from a high-performance search engine, this book aims to provide a comprehensive and novel account of the stress-assignment properties, selection processes, productivity and combinatorial restrictions of native and non-native suffixes in Present-Day English. In a resolutely interscholastic approach, the author has confronted his findings with the tenets of Generative Phonology, Cyclic Phonology, Lexical Phonology, The Latinate Constraint, Base-Driven Lexical Stratification, Complexity-Based Ordering and Optimality Theory.

Table Of Contents

  • Cover
  • Title
  • Copyright
  • About the author
  • About the book
  • This eBook can be cited
  • Contents
  • Symbols and conventions
  • Abbreviations
  • 0. Introduction
  • 0.1 Objectives and methodology
  • 0.2 Stress-assignment: a confrontation between two phonologies
  • 0.3 Two families of affixes to account for the combinatorial properties of affixes?
  • 0.4 Rules vs. constraints
  • 0.5 Book structure
  • Part I. S-1 and auto-stressed suffixes
  • 1. -ic
  • 1.1 General features
  • 1.2 Suffix combinations
  • 1.3 Allomorphic transformations
  • 1.4 Extensions of the -ic rule
  • 1.5 Summary and conclusion
  • 2. -ion and similar affixes
  • 2.1 General features
  • 2.2 The -ION generalisation
  • 2.3 -ion and its allomorphs -ation, -ition, -ution, -fication, -faction
  • 3. -ity
  • 3.1 General features
  • 3.2 Suffix juxtaposition and substitution
  • 3.3 -ety
  • 3.4 -ty
  • 3.5 Underived nouns in -ity
  • 3.6 Summary and conclusion
  • 4. S-1 suffixes indicative of smaller word populations
  • 4.1 -ify
  • 4.2 -ible/-igible
  • 4.3. -icide
  • 4.4. -meter
  • 4.5 -erie
  • 5. Stress-bearing affixes
  • 5.1 Affixes of French origin
  • 5.2 Stress-bearing affixes from Latin or Romance languages other than French99
  • 5.3 Germanic stress-bearing suffixes
  • 5.4 Neoclassical affixes and combining forms
  • Part II. Neutral suffixes
  • 6. Grammatical suffixes
  • 7. Consonant-initial suffixes
  • 7.1 General features
  • 7.2 Consonant-initial suffixes of Latinate stock in Present-Day English
  • 7.3 Consonant-initial suffixes of Germanic stock
  • 7.4 Consonant-initial suffixes of Germanic stock still productive in Present-Day English
  • 7.5 Suffixes extracted from foreign words
  • 7.6 Neoclassical combining forms
  • 7.7 Summary and conclusion
  • 8. Neutral vowel-initial suffixes of Germanic stock or of uncertain origins
  • 8.1 Unproductive forms
  • 8.2 Productive forms
  • 9. -er
  • 9.1 General features
  • 9.2 Productivity in compounds
  • 9.3 Productivity in non-compound lexemes
  • 9.4 Nouns in -er with an obscure or opaque stem
  • 9.5 Suffix stacking
  • 10. Latinate vowel-initial suffixes: -er’s rival agent noun suffixes
  • 10.1 -ant/-ent
  • 10.2 -ator and -or
  • 10.3 -ist
  • 10.4 -ite
  • 10.5 Unproductive person or instrument suffixes
  • 10.6 Summary and conclusion
  • 11. Latinate Vowel-initial noun suffixes of action, state, process and result
  • 11.1 -acy
  • 11.2 -age
  • 11.3 -al
  • 11.4 -ance/-ancy, -ence/-ency
  • 11.5 -ate
  • 11.6 -ery
  • 11.7 -ule
  • 11.8 -ure
  • 11.9 -Mixed suffixes
  • Part III. Mixed suffixes
  • 12. -able
  • 12.1 General features
  • 12.2 -able or -ible?
  • 12.3 Stress-neutrality and variation
  • 12.4 Suffix stacking
  • 13. Verb suffixes
  • 13.1 -ate
  • 13.2 -ise
  • 14. -y and -ism
  • 14.1 -y
  • 14.2 -ism
  • Part IV. S-1/2 suffixes
  • 15. Adjective suffixes
  • 15.1 #Syl + -al, -an, etc
  • 15.2 -ION adjective affixes
  • 15.3 Consonant clusters + adjective affixes -al, -ous, etc
  • 15.4 Vowel digraphs + -al, -an, etc
  • 15.5 -ul- + adjective affixes -ar, -an, -ous, etc
  • 15.6 -VCal/-an/-ous, etc
  • 15.7 -ative, -atory, -utive, -utory
  • 15.8 Suffix stacking
  • 16. Neoclassical suffixes
  • 16.1 General features and stress assignment
  • 16.2 Productive suffixes
  • 16.3 Exceptions to truncation of neoclassical endings
  • 17. Stress-assignment and suffix stacking, overall recapitulation431
  • 17.1 Stress-assignement
  • 17.2 Suffix stacking
  • Part V. Further issues
  • 18. Compounds
  • 18.1 Combining-form compounds
  • 18.2 Standard compounds
  • 19. Conversion
  • 19.1 Noun-verb and verb-noun conversion
  • 19.2 Adjective-noun and noun-adjective conversion
  • 19.3 Adjective/verb conversion
  • 19.4 Verb-adjective conversion
  • 20. Secondary stress
  • 20.1 General principles
  • 20.2 The condensation/information dichotomy
  • General conclusion
  • References

Symbols and conventions

C consonant
C2 consonant cluster (at least two graphic and/or phonological consosants)
V vowel
VDig vowel digraph
< > graphic notation
/ / phonological notation
[ ] phonetic notation
# morpheme boundary
/ derivational alternative (cubism/cubist < cube) or variational or synonymous pair separator: anticipative/atory
+ morphological component boundary (de- + material + -is(e) + -ation) or bound or stress-placing affix in Lexical Phonology literature
< historically derived from: boundary (< bound + -ary)
> reverse-order approach to historical derivation: bound > boundary
<~ synchronically derivable from: rejection (15th < Latin) but synchronically parseable as derived from reject further to attachment of the suffix -ion (rejection <~ reject)
~> reverse-order approach to synchronic derivational patterns: reject ~> rejection
<≠ is not derived (or synchronically not derivable) from: ignorant (“lacking in knowledge or unaware”) <≠ ignore (“refrain from noticing or acknowledging”)
≠> reverse-order approach to semantic demotivation: ignore ≠> ignorant
 different from or not synonymous with
* ungrammatical or unattested form: *plentifulise
?? unattested but potentially licit form: ??problemsome ← xiii | xiv →


← xiv | xv →


1. General usage

act. actually Gram. grammar
adj. adjective insep. inseparable
adv. adverb lang. language
alt. alternative(ly) Math. mathematics
arch. archaic n. noun
BF back-formation norm. normative
bef. before obs. obsolete
CF combining form orig. origin or originally
ch. chapter par. paragraph
cont. contemporary pl. plural
cp. compare prob. probably
D. Dictionary (e.g. D. com for
Dictionary. com)
rel. to
relating or relative to
decomp. decomposable resp. respectively
def. definition s. sense(s)
dem. demotivated sep. separable
der. derived or derivable sim. similarly
dial. dialectal spec. special or specialised
diff. different syl. syllable(s)
dim diminutive syn. synonym(ous)
esp. especially sync. synchronically
etym. etymology ult. ultimately
exc. exception(s) v. verb
fig. figurative Zool. Zoology
freq. frequentative  

← xv | xvi →


2. Languages

A Ancient
Alg. Algonquin
Ar. Arabic
Chin. Chinese
Da. Danish
Du. Dutch
E English
F French
G German
Gmc Germanic
GB British English
Gk Greek
H High
Heb. Hebrew
Hin. Hindi
Ir. Irish
It. Italian
Jap. Japanese
L Latin (or “Low”, as in MLG = Middle Low German)
Med. Medieval
M Middle
Nor. F. Norman French
O Old
Per. Persian
Por. Portuguese
Rus. Russian
Sc. Scots
Scan. Scandinavian
Sp. Spanish
Sw. Swedish
US American English ← xvi | xvi →

0. Introduction

0.1 Objectives and methodology

The aim of this book is to provide a comprehensive assessment of the role of suffixes in lexical stress-assignment and word-formation, complete with a systematic overview of their selection processes, productivity and combinatorial properties in Present-Day English.

A methodological prerequisite which has become incontrovertible in language studies is the necessity to draw upon a reliable corpus. The multiplication of online databases has provided researchers with worktools many times more powerful than those they had at their disposal not so long ago. The corpus used in the present study has been assembled from the OneLook search engine (henceforth OL) which, in English, enables users to extract word inventories further to a preselection of morphological components from about a hundred generalist or specialist dictionaries1.

So as to warrant indisputable reliability as to the data exploited, the corpus used in this study has been culled from the entries of seven generalist dictionaries whose reputation is solidly established, complemented with those of Dictionary.com which is the only OL dictionary providing full etymological data in most of its entries2.

The dictionaries from which the OL corpus has been established are, by alphabetical order: ← 1 | 2 →

(1) (→ = henceforth)
American Heritage Dictionary of the English language (→ American Heritage D.); Cambridge International Dictionary of English (→ Cambridge D.); Collins English D. (→ Collins D.); Dictionary.com (→ D.com); Encarta World English Dictionary, North American Edition (→ Encarta D.)3, Merriam-Webster’s Online Dictionary, 11th Edition (→ MWD); Webster’s Revised Unabridged, 1913 edition (→ Webster’s D.); Wordsmyth English Dictionary/ Thesaurus (→ Wordsmyth D.).

The corpus assembled from the dictionaries in (1) contains 140,000 words.

Common words recorded exclusively in the Wikipedia Encyclopedia and its companion dictionary Wiktionary have not been retained in the corpus selected (henceforth the Corpus) since they do not meet the reliability criteria prescribed in lexicography. References have however been made to these online databases, notably to measure the potential productivity of some highly specialised or strictly scientific families of words, since Wikipedia and Wiktionary have obtained such items from scientific literature. For example, the combining form -saurus (< New Latin) is found in only a score of learned compound in the Corpus (e.g. Apatosaurus, Megalosaurus) vs. nearly 1,000 nouns of extinct saurian species identified by palaeontologists which have hitherto been recorded only in Wikipedia.

Although, because of the nature of the corpus used in this study, hapaxes have not been made a priority criterion in productivity measurements, well-formed words (especially neologisms) obtained from the Internet have regularly been sampled when they did not appear in the Corpus, with the policy of retaining only those which occurred in high-register Web pages (e.g. scholarly texts, official documents). Exception to this vetting process has only been made when there was the necessity to deal with new suffixes used in recent media (computing, internaut fora, video games, etc.). Such cases have been scrupulously pointed out.

Lexical and variational information about the additional items gathered from Web Pages has been verified from other online databases, namely: A Cross Reference of Latin and Greek elements, Dictionary/Thesaurus, ← 2 | 3 → Dinosaur/Palaeontology Dictionary, findtheword.info, Free Dictionary, Infoplease Dictionary, Macmillan Dictionary, Medical Dictionary, Memefirst, Online Medical Dictionary, Rice University Neologisms Database, Urban Dictionary, Wordnik, World Wide Words, The Word Spy4, (cf. References for Web addresses).

Measuring the productivity of affixes from the data returned by dictionary-based corpora can, naturally, not consist in merely considering the number of words they have yielded, such inventory counts being only relevant in a historical perspective. As will be shown below a fair number of suffixes which have played a major role in enriching the lexicon are now extinct or obviously in their death throes. Focusing on recently-formed words, roughly over a period which can be placed from the 20th century – more particularly after World War II, which saw a spectacular development of mass-media diffusion and interdialectal exchanges – to the present day, seems to be an effective and relevant protocol, provided due attention is paid to the fact that affixes are inevitably subject to fashion trends, which implies that a recently-adopted suffix may be ephemerally successful. Conversely, a suffix may suddenly rise from its ashes after years of obsolescence (cf. -age).

It is at this stage important to remind the reader that synchronically transparent derivational sequences are not necessarily faithful reflections of the history of English. Many putative suffixed formations have actually been directly borrowed from French, Latin or Ancient Greek or tardily reconstructed on roots from the last two languages. Such items have been signalised as follows: concision (14th < L <~ concis(e) + -ion), to be read as “concision, adopted from Latin in the 14th century, a noun synchronically analysable as resulting from the affixation of -ion to the adjective concise”.

Whereas D.com is, as said above, the only dictionary of the Corpus providing full etymological data for most of its entries, many potential derivatives are given in this dictionary merely as “related forms”, without further information, e.g. applause (main entry, with phonetic transcription, syntactic category, definitions and etymology) and applausive (tagged as a “related form” below applause, with stress pattern ← 3 | 4 → and syntactic category, but no specification as to whether this adjective was formed from the noun above, with attachment of the -ive suffix, or directly adopted from French or Latin). Moreover, as regards both the origins of words and their dates of earliest known use, discordant data are rife amongst dictionaries, a fact which is particularly striking when comparing D.com’s etymological notices with those of the online edition of The Oxford English Dictionary (<www.oed.com>, henceforth OED), which is not accessible from the OL search engine.

Still with the concern of warranting maximum reliability as to the data exploited, the dates of earliest occurrence of words given in further pages have been taken from OED, which has the incomparable advantage of providing written sources for this purpose. Dates of earliest attestation provided by D.com, standardly given in the form of 5- or 10-year time frames (e.g. sportster 1960–65), have not been retained since they often differ widely from those indicated in OED (e.g. sportster 1918). In the few cases where no information on the earliest known use of a word could be obtained from OED, relevant data have been garnered from MWD. For reading convenience, dates of earliest known use indicated in samples or inventories have been given in the form of centuries except for words which are to be considered as recent according to the criteria set out two paragraphs above, namely words which first appeared in the 20th century (e.g. phenomenal 19th instead of 1825 vs. rapster 1981). This convention has been breached when need arose to clarify which item of a paradigmatic set was first attested in English (e.g. nominal pairs in -pathy and -path: psychopathy (1847) / psychopath (1864)). Dates of earliest appearance given in further pages are thus compliant to the following conventions:

  • a. pheneomenal (19th); psychopathy (1847) / psychopath (1864); funkster (1963); no other indication than date = date of earliest known use obtained from OED
  • b. injectant (1950, MWD) = word not listed in OED, date of earliest known use obtained from MWD

As for the origins of items exemplified in further pages, authentic derivations, given with the < or > symbols, reflect either etymological concordance in D.com and in OED, or etymology from OED alone for want of relevant information in D.com, ie items tagged in the latter dictionary ← 4 | 5 → as “related forms” with no further specification. Thus, examples such as categorise (< categor(y) + -ise) and criticisable (< criticis(e) + -able) have not been differentiated in further samples and inventories, although the first item is given the same etymology in D.com and in OED whereas the second is merely labelled as a related form of criticise in D.com vs. as derived from criticise in OED. Conversely, disagreements between both dictionaries as to words which are synchronically analysable as formed by derivation have been systematically pointed out as follows: criminalistics (1910 < criminalist + -ics vs. < criminalistic + -s in OED), to be read as “date of earliest known use from OED (cf. (2) above), followed by (reference omitted for reading convenience in this and further instances) D.com’s derivational description vs. OED’s”); other example: instantiate (1949 < L stem + -ate vs. < instanc(e) + -iate in OED). Finally, The Online Etymology Dictionary (henceforth OEtymD), which is accessible from OL, has occasionally been appealed to, especially with regards to the history of some suffixes.

In this book, semantically transparent deriving forms have been referred to as bases, the term stem being reserved for bound and opaque morphemes further to the removal of an inseparable affix (e.g. *patern in paternal or *joice in rejoice).

Issues at the centre of this study have elicited the interest of linguists claiming adherence to diverse schools and disciplines, among which morphophonology, morphosyntax, morphosemantics and psycho-linguistics figure prominently. Among the theoretical avenues which have underlain research in affix properties, morphophonology has long been the most popular given the role played by affixes in lexical stress-assignment. One of the most spectacular recent developments in the study of English affixes has led its proponents (cf. §0.3 below) to postulate usage restrictions stemming not only from affixes but from base types.

Although it initially set out to depart radically from former theoretical frameworks, most notably Cycle Phonology, Optimality Theory has brought forth a new generation of researchers (e.g. S. Collie, R. Raffelsiefen, see references in further chapters) who have attempted to incorporate into their work what they deemed still relevant from previous leading theories. ← 5 | 6 →

Such efforts are to be highly commended as interscholastic research work in linguistics is often still met with a disapproving eye. Yet, whatever their respective merits, the models which have emerged over the last decades would gain greatly by opening up to alternative theoretical views, the issue of the interplay between morphology and phonology leaving still too many open questions to be locked into pigeon-holed orthodoxy.

Openness to various approaches is precisely what has underlain the conception of the present work, with no concessions for erroneous conclusions past or present models may have generated but respectful consideration for the legacy accrued by such eminent linguists as Aronoff, Bauer, Burzio, Chomsky, Fabb, Fudge, Giegerich, Guierre, Halle, Hay, Keyser, Kiparsky, McMahon, Marchand, Mohanan, Plag, Poldauf, Prince, Siegel, Smolensky, Spencer and Vergnaud, to name just a few.


ISBN (Softcover)
Publication date
2015 (January)
Morphophonology Linguistics Noam Chomsky Affix
Bern, Berlin, Bruxelles, Frankfurt am Main, New York, Oxford, Wien, 2015. 471 pp.

Biographical notes

Ives Trevian (Author)

Ives Trevian is a tenured senior lecturer accredited to direct doctoral research in Linguistics at Paris-Diderot. His publications – which include two books published by Peter Lang in 2003 and 2010 – have centred on stress assignment, morphophonology, affixation processes, neoclassical compounds and English-language history.


Title: English suffixes