ChapterPDF Available


This is an overview of morphological means employed in constructions of comparison and gradation, composed for the HSK manual of morphology.
Christiani Lehmanni inedita, publicanda, publicata
Pierluigi Cuzzolin coauctore
Comparison and gradation
huius textus situs retis mundialis cuzzolin_lehmann_comparison .pdf
dies manuscripti postremum modificati
occasio orationis habitae
volumen publicationem continens
Booij, Geert & Lehmann, Christian & Mugdan, Joachim &
Skopeteas, Stavros (eds.), Morphologie. Ein Handbuch zur
Flexion und Wortbildung. 2. Halbband. Berlin & New York:
W. de Gruyter (Handbücher der Sprach- und
Kommunikationswissenschaft, 17/2)
annus publicationis
Comparison and gradation
1. Preliminary notions
2. Comparative
3. Superlative
4. Equative
5. References
1. Preliminary notions
All the languages of the world have at their disposal different means to express comparison
and gradation, but not every language expresses them through morphology. In recent years
some relevant works on the typology of comparison were published (Stassen 1985;
Xerberman 1999; but see already Jensen 1934), and the semantics of comparative
constructions is being investigated by semanticists. A reliable survey on comparison and
gradation, however, still needs detailed research in several theoretical domains, including
morphology (see, for instance, the short entries by Andersen 1992 and Crookston 1994).
In the present article, then, comparison and gradation will be treated to the extent that
morphology is involved, and a classification of the means employed for its expression will be
Viewed conceptually, both comparison and gradation presuppose an entity that some
property, state or, more rarely, a more or less dynamic state of affairs applies to. They also
presuppose that this state of affairs varies on some scale on which it may, in principle, be
measured; i.e. it is a parameter. Gradation (German Abstufung) then is the stepwise
modification of the extent to which the parameter applies to the entity, while comparison
(German Vergleich) assesses this extent with respect to some standard. Taken as a
grammatical category, comparison (German Steigerung) is the formal modification of some
predicative word – most often an adjective – representing a parameter of gradation or
comparison, according to the extent to which it applies to its argument, relative to some
standard. Similarly, gradation may be manifested in a structural category, in particular of
adjectives and verbs.
In a comparative construction four elements are identifiable:
(a) the element which is compared, the comparee, or topic of comparison;
(b) the element that serves as standard of comparison;
(c) the element, called the pivot, or marker of comparison, that introduces the standard
of comparison;
(d) the predicate that represents the parameter of comparison.
(1) John is smart-er than Sam
comparee predicate-CMPR pivot standard
Elements (b) and (c) can be absent for different reasons:
- the pivot, as some languages possess no element for that function (see 2.1). The
construction of (1) is in fact very rare among the languages of the world, and the standard is
more frequently marked by other morphological means, for instance case markers;
- the standard of comparison can be omitted for semantic and pragmatic reasons. In fact,
although grading involves comparison (Sapir 1951; Lyons 1977:271), the element or the
entity serving as the standard can be either presupposed logically (see, for instance, Panagl
1973 and Bertocchi & Orlandini 1997 for Latin) or by conversational implicature (see
Chierchia & McConnell-Ginet 1990:283-285). Sentences like I thought he lived in a cleaner
house or Sheila is less lucky are rather common in discourse. Comparatives occurring without
standard, traditionally labelled absolute comparatives seem much more frequent than those
occurring with the standard, but statistics on this point are scarce.
The adjective is the lexical category that typically undergoes comparison. Three
degrees are traditionally recognised: positive, comparative, and superlative.
The positive degree (from Lat. positivus 'imposing (the standard)'), coincides with the basic
form of the adjective itself, e.g. Engl. a small box, an interesting book.
The comparative degree of some predicate – typically an adjective – marks this
predicate as applying to its argument (the comparee) to a higher extent than the standard; e.g.
smaller. In gradable predicates, esp. in polar (or relative) adjectives, the form that is
unmarked for gradation or comparison – for adjectives: the positive degree – semantically
involves implicit comparison with a norm. For instance, an old dog is a dog that is older than
some standard for dogs. It is the function of the comparative degree to allow for the
substitution of the implicit default standard of comparison by a particular explicit standard of
From a cognitive point of view, it is more salient to describe entities that are larger, smaller,
than those that are less large, less small and so forth. In grammars the former type is referred
to as comparison of majority, the latter as comparison of minority: That box is smaller
than this vs. Your book is less interesting than hers. Whereas the comparison of majority is
expressed by morphological processes in several languages, apparently the comparison of
minority is expressed only lexically. This implies that there is no affix meaning 'less' parallel
to that of majority 'more' (but see 2.4).
A comparison of equality is one that ascribes to the comparee the same value of the
parameter of comparison as to the standard. If it is marked formally on the predicate
representing the parameter – typically, an adjective – this is called the equative form (see 4).
In order to describe this construction, five elements have been identified (Haspelmath
& Buchholz 1998:279):
(2) My sister is as pretty as you
1 2 3 4 5
where 1 is the comparee, 2 is the parameter marker, 3 is the parameter, 4 is the standard
marker, 5 is the standard.
Very close to the equative is the similative, "a construction expressing sameness of
manner" (Haspelmath & Buchholz 1998:278): Robert is as tall as Maria is equative, whereas
She sings like a nightingale is similative. Similative constructions will not be treated here.
By the superlative (Lat. superlativus 'raised to the top'), the quality expressed by the
adjective is described as being at a very high, possibly the highest degree. However, it must
be observed that "the superlative form is often used to denote a high grade, but not necessarily
an apical grade, of the graded quality" (Sapir 1951:146).
In some languages (Classical Arabic, for example), the adjective shows one single
form that exerts the function both of the comparative and superlative, without any
morphological distinction; this special form is frequently referred to as elative (Latin elativus
'standing out'): akbar, from kabi:r, for instance, means 'greater, greatest, very great' according
to the syntactic environment or the context.
The superlative degree in comparison (e.g. Zembo is the laziest of all the chimps at the
zoo) is to be distinguished from the highest degree in gradation (e.g. a very old tree, the
easiest way to clean). Languages like English use different forms. The former is expressed by
the superlative form of the adjective (the positive form either preceded by most or followed by
-est), generally preceded by the definite article: the most beautiful, the luckiest. The latter is
marked by an intensifier preceding the positive adjective: very lucky. Forms like *the very
lucky of all men are ill-formed. Languages like Latin use the superlative in both cases: a
superlative like felicissimus, for example, occurs both in gradation (vir felicissimus 'man very
lucky') and in comparison (vir omnium felicissimus 'the luckiest man of all'). The two uses of
the superlative form are then called absolute and relative superlative, respectively.
The following points concern formal properties of the structural categories of
comparison and gradation. First, the status of comparison between derivation and inflection is
controversial and may differ according to the language. Second, suppletive expression (see
Art. 52) is not rare: Latin bonus 'good', melior 'better', optimus 'best'; Hung. sok 'good', több
'better'. While analytic expression of gradation (e.g. very tired) belongs in the realm of syntax,
it may also be expressed by compounding (Dutch dood-moe (dead-tired) 'very tired') and
affixing (Latin per-bonus 'very good', per-terrere 'frighten very much').
Third, while comparison of adjectives and (derived) adverbs is well-known, it does
occur on nouns, too. Cf. Sanskrit vira-tará 'more hero', Ancient Greek kún-teros '(more)
shameless', lit. 'more dog'; Hungarian róka 'fox', comparative rókább 'slyer' (Bhat 1994:25f.).
In Basque the suffix -(a)go can be added to any lexical category: gizon 'man': gizonago 'more
man'; but also gugana 'towards us': guganago 'more towards us'. In Italian the superlative
suffix -issim- is added to nominal stems to express the highest degree of a quality:
governissimo 'a stable and powerful government', from governo 'government'; canzonissima
'the best song (among those in competition for an award)', from canzone 'song'.
Fourth, languages may have semantic, morphological or phonological constraints on the
application of comparison to members of a word class or on its expression. The alternation
between synthetic and analytic comparison in English is due to a formal constraint. Most of
the semantic constraints result from the defining criterion of comparison which essentially
involves a gradable parameter. Consequently, adjectives with complementary (contradictory)
meaning such as odd/even are not used in comparison and gradation, at least not in their literal
sense, and therefore often lack the morphological category of comparison. Also, certain
adjectives that designate the highest grade tend to avoid comparison and gradation. This is
true both for explicitly derived forms such as Dutch doodmoe, *doodmoeër/*doodmoest
(Booij 1996:5) and for lexical terms for the highest degree such as excellent, *very excellent.
2. Comparative
According to the absence or the presence, either optional or obligatory, of a marker of
comparison on the lexical category involved, typically adjectives and adverbs (one of the
criteria advocated by Xerberman 1999; but see already Jensen 1934), we can subdivide the
strategies into four main groups.
2.1. No marker
The graded form has no marker. One of the most widespread types of comparative structures
is the juxtaposition of two positive adjectives, with no marker of comparison and with
antonymic value: X (is) long, Y (is) short = X is longer than Y. This is very common in
isolating languages and can be found also among languages with agglutinative or fusional
morphology. As Benveniste (1948) had already suggested, this strategy seems to be basic and
independent of the morphological type of language in which it occurs:
"La comparaison est implicite dans l'énoncé antithétique. Toutes les langues peuvent recourir à ce moyen, même
quand elles disposent d'une expression morphologique." (Benveniste 1948:126)
This type is illustrated by Samoan in (3) and occurs very frequently among
Amazonian and Austronesian languages.
(3) Ua loa lenei va'a, ua puupuu lena.
is long this boat, is short that
'This boat is longer than that.'
A subtype is one in which a positive adjective is juxtaposed with the negated form of
the same adjective: i.e. X (is) long, Y (is) not long = X is longer than Y, like in Apalai (Koehn
& Koehn 1986:52):
(4) Mopo zumo pyra kyn-exi-ne akono zumo.
Mopo big NEG 3-be-DP big
'His brother-in-law is bigger than Mopo.'
This strategy has been reconstructed for Proto-Indo-European (Puhvel 1973) and
is also attested in Vedic and the Baltic languages (Pinault 1985; 1989), possibly in Latin
(Baldi & Cuzzolin 2001:220). The two subtypes mentioned often cooccur in the same
language, like, for instance, in Wari' (Everett & Kern 1997:193f).
2.2. Optional lexical morpheme
The form is graded by means of an often optional lexical morpheme, mainly unbound. In this
type, too, which is very frequent among the languages of the world, the adjective has no
marker of comparison and the standard occurs with a case marker, generally coinciding with
some case (genitive, ablative, dative are the most frequent). As far as known, no language is
attested in which the standard of comparison is marked by a special case, although in some
languages (for instance Limbu (van Driem 1987:52) or Tsaxur (Sosenskaja 1999:561)) the
standard is formed by a combination of two, mainly semantic cases that form a "special
This type is rather widespread particularly among the agglutinative languages: in
Turkish either the positive adjective alone or with the adverb daha is currently employed
(Lewis 1988:54):
(5) kursun-dan (daha) agır
lead-ABL (more) heavy
'heavier than lead'
In Udihe, a Manchu-Tunguz language, "the adjective itself does not take any special
comparative morpheme" (Nikolayeva & Tolstaya 2001:180) and the standard is in the ablative
(6) Ussuri Biki-digi su
Ussuri Bikin-ABL deep
'The Ussuri is deeper than the Bikin.'
In Japanese the adjective is normally expressed without any adverb: Tiroo wa Taroo
yori kasikoi desu 'Chiroo TOP Taroo from smart is (Chiroo is smarter than Taroo)' but the
emphatic form no hoo ga 'very' can be added after yori, without any basic change in the
meaning of the sentence (Hinds 1986:127f.).
The adverb meaning 'more' and the pivotal element corresponding to than may be
agglutinated into one phonological word: Tümpisa Shoshone (Uto-Aztecan):
(7) Üü yuhupi wakakwa.
you fat me more:than
'You are fatter than me.' (Dayley 1989:288f.)
2.3. Affix
The form is graded by means of an affix. As has been repeatedly noticed, this type, which is
quite common in the Indo-European family, is extremely rare among the world's languages.
English employs two morphological processes to form the comparative: in smaller
there are the lexical morph small- and the suffix -er, which conveys the meaning 'more',
whereas in more interesting the notion of comparison is expressed lexically by the adverb
more. In the former case the comparative is expressed synthetically, in the latter analytically.
This is in keeping with the general tendency observable among different linguistic families to
replace synthetic by analytic expression. In English the use of the suffix -er is limited by
phonological constraints (Quirk et al. 1994:461f.); it is no longer productive and tends to be
replaced by the adverb more: smaller/more small.
Other declension categories may interact with the comparative in various ways. The
simplest subtype is represented by a language like Old Irish, in which "all the forms of
comparison are uninflected, and show no difference of number or gender" (Thurneysen
1980:232). For instance, the adjective dían 'swift' has the comparative déniu 'swifter' for all
the numbers and genders. In Latin, oblique case forms of comparative adjectives contain an
agglutinative sequence of the comparative suffix plus a case/number suffix which neutralizes
gender: beat-ior-i 'happy-CMPR-DAT.SG', whereas in the so called direct cases (nominative,
accusative, and vocative) there is a difference between beat-ior-em 'happy-CMPR-SG.M/F.ACC'
and beat-ius 'happy-SG.N.ACC': in the latter the marker of comparison also conveys the
grammatical information of number, gender and case, with cumulative exponence (cf. Art.
There are two other very rare morphological processes to express gradation. Among
the Semitic languages degree forms exhibit a stem with structure /aCCaC/, derived from the
normal positive form of the adjective, usually /CaCCi:C/; e.g., Arabic hami:d 'praiseworthy',
elative ahmad 'more praiseworthy'.
Among the Kartvelian languages the comparative can be formed by circumfixes: in
Georgian, the comparative is characterised by the circumfix u-…-es with possible vocalic
variation in the adjectival stem: lamezi 'beautiful', comparative u-lamez-es-i 'more beautiful';
magari 'strong', u-magr-es-i 'stronger' (Hewitt 1995:48). This type is no longer productive in
Modern Georgian. In Svan, too, the comparative is formed by the circumfix xo--a and the
adjective can also show vocalic variation: c’rni 'red', xo-c’ran-a 'redder' (Tuite 1997:18).
Finally, an interesting case is represented by Modern Irish:
"There is only one degree of morphologically expressed comparison... It is formed from the base adjective
usually by palatalisation of the final consonant and the addition of -e..." (O'Dochartaigh 1992:74).
This form, however, is preceded by the particle níos: bán 'white', níos báine (ná) 'whiter
(than)'. In Modern Irish, therefore, the comparative form is marked twice, lexically and
2.4. Verb
The adjectival category, which clearly exhibits the morphological features typical for nouns in
the Indo-European languages, is more verb-like in numerous other languages (a useful
analysis in Bhat 1994:187-209). This phenomenon is particularly frequent among isolating
languages, or those showing a low degree of synthesis, and among the native languages of
America (but see Bhat 1994:211-243). We refer to this type as the "verbal type of
comparison". Within this type, two subtypes are identifiable. The first involves verbs meaning
'to surpass, to overtake'. Needless to say, the expression of comparison here involves syntax
rather than morphology. In the second subtype, the parameter of comparison itself is a verb.
In Saek, a language of the Austro-Thai family spoken between Laos and Thailand,
comparison is expressed by the verb lyyn5 'surpass, overtake' (Morev 1988:43; numbers in
exponence refer to tones):
(8) lum4 myy3nii5 reeng4 lyyn5 myy3luan4
wind today strong surpass yesterday
'today's wind is stronger than yesterday's'
This type is also attested in other South-East Asian languages, genetically related like Zhuang
(Moskalev 1971:198) or unrelated, like Vietnamese (Nguyên 1997:122; see also Stassen
1985), but it is not limited to isolating languages: it also occurs in agglutinative or fusional
languages with verbs meaning 'to overtake, to exceed'. It is widespread among the Bantu
languages and also occurs among the Chadic languages: Hausa (cf. Sceglov 1970:239, 273;
Miya, cf. Schuh 1998:314). In Rwanda, for instance, "to express comparison between two
qualities or attributes the verbs kurusha and kuruta 'exceed, surpass' are used, which agree in
person, number and class with the nouns denoting a person or thing that surpasses another
person or thing participating in comparison" (Dubnova 1984:45).
In Amharic (Afro-Asiatic), the verbs läqqa 'surpass' or bällätä 'exceed' can reinforce
the comparative form of the adjective (Hudson 1997:466):
(9) Haylu k-antä ybält
Haylu from-2.SG.M exceed
bätam qäccn nä-w.
very thin is-he
'Haylu is much thinner than you.'
This type of comparative, which according to Stassen (1985:159; on the relationship
between comparison and word order see also Andersen 1983 and Romero-Figueroa 1986)
tends to be related to SVO languages and supposedly developed in Late Proto-Indo-European
(Puhvel 1973), also occurs in languages where the commonest type is the juxtaposed type as
in Wari' (Everett & Kern 1997: 194f.).
In some languages a similar structure is used to express the comparative of minority,
employing a lexical item meaning 'behind', cf. (9) from Miskito (Misulmapan) (Suárez
(10) witin jang ninara tukta
he I behind child
'he is younger than I'
A more appropriate translation of (10) would be 'he is less old than I'. A parallel construction
is found in Amazonian languages like Sanuma (Borgman 1990:54).
The comparative construction necessarily has a verbal head in those languages in
which property concepts are verbs, such as West Greenlandic (Greenlandic Eskimo):
"Comparative degree is expressed derivationally with a comparative/superlative affix on the
(verbal) base expressing the parameter of comparison and, optionally, a case-marked nominal
expressing the standard or limit of comparison (this may also be a possessive inflection on the
comparative morpheme when it is in nominal form)." (Fortescue 1984:167f.):
(11) kujataa-nit issin-niru-vuq
south-ABL be.cold-more-3.SG.IND
'it is colder than the south'
"In the older language niru could be left out in fact" (Fortescue 1984:168).
2.5. Diachronic tendencies
Some tendencies are observable in the diachrony of comparative constructions. First of all,
some old synthetic superlatives become opaque forms and tend to be felt as positive
adjectives, for which new comparative and superlative forms are created. In Italian for
instance, intimo has nowadays almost completely lost its original superlative function of
'innermost' and only means 'intimate', with a comparative (più intimo) and a rather rare
intimissimo, used in commercials. But even in Latin, especially Late Latin, we find that those
superlative forms which were residual and built by means of non-productive rules, tended to
be reanalyzed as positive: a form like extremius (recorded since 2nd c. A.D.) is a comparative
formed from the superlative extremum 'very far, extreme'.
This phenomenon should not be confused with the tendency to reinforce and
strengthen comparative forms, a feature which is rather common among the Indo-European
languages. For instance, the co-occurrence of two markers of comparison was rather frequent
in Late Latin: forms like magis fortior = fortior/magis fortis 'stronger', magis beatior =
beatior/magis beatus 'happier' are well attested (Hofmann & Szantyr 1965:166f.). This
phenomenon also occurs in spoken varieties of fusional languages which still retain the
synthetic form of the comparative beside the analytic one: see Modern Greek pio mikróteros
'more smaller' (Holton et al. 1997:87).
Secondly, a well-known phenomenon in the history of numerous languages is the
replacement of synthetic forms by analytical ones. This is largely attested in the history of the
Romance languages, where the synthetic forms in -iore(m) were gradually but nearly
completely replaced by the analytical forms deriving from plus (Italian più and its dialects,
French plus; Modern Greek pio, as in pio kaló 'more beautiful', ultimately derives from plus
via Venetian) or from magis (Spanish mas: mas hermosa 'more beautiful', Portuguese mais:
mais famoso 'more famous', Rumanian mai: mai înalt 'higher'), and in several spoken varieties
around the Mediterranean.
In Tunisian Arabic the form of the elative (template /aCCaC/: hasan 'good', ahsan
'better, very good') is usually replaced by the form aktar (elative form of kati:r 'much') plus
the positive: aktr hasn 'better'.
3. Superlative
The remarks in 2 concerning comparatives hold also for the superlative. It is worth noting
that, although superlatives are often formed by the same morphological process as
comparatives, they also show some typical formations. There are two basic, non-exclusive
morphological processes used by the languages of the world. They are, however, mainly
employed to form the highest grade (or absolute superlative), whereas the relative superlative
is basically formed by syntactic devices.
3.1. Analytic formation
The form is graded by means of a lexeme. This is probably the commonest means of
expressing the highest grade among the world's languages: almost every language has a word
meaning roughly very which, preposed or postposed, combines with the adjective (see Klein
1998 on the semantics of adverbs of degree): very friendly, Xhosa inencasa gqitha 'delicious
very (too delicious)'. In several languages this is the only possible means creating the
For stylistic, expressive reasons particular adverbs are employed in discourse: e.g.
English terribly, German furchtbar 'terribly', Italian straordinariamente 'extraordinarily'
(Austerlitz 1991:3). In English terribly good simply means 'very good'. Adverbs like those
mentioned are often the etymological source of the adverb 'very': very itself originally meant
'really'); German sehr, which is the normal adverb for 'very' in present-day German, is the
grammaticalised Middle High German adverb sere which in origin meant 'painful, violent'.
3.2. Synthetic formation
The adjective is graded by means of some morphological process. In this case the inventory of
the morphological processes displayed among the languages of the world is also quite rich.
One of the processes by which the superlative is formed is reduplication (cf. Art. 57;
extremely rarely used in the formation of comparative, as in Upriver Halkomelem; see
Galloway 1984:56). In some cases there is total reduplication of the adjective, as in Samoan:
tele 'big', teletele 'very big' or Sumerian: bar 'external', barbar 'very external = foreigner,
barbarian'. In colloquial Italian one can reduplicate the adjective with superlative function in
predicative position: È rimasto lí tranquillo tranquillo 'He remained there very quiet', but with
an evaluative nuance.
Other cases involve partial reduplication like, for instance, of the first syllable (with
some change) in some Mongol languages. In Modern Mongolian, for example, the first
segment /(C)VC/ of the adjective is reduplicated but the last phoneme is replaced by /v/, so
that /(C)VC/ becomes /(C)Vv/ (inexact reduplication: Art. 57): ulaan 'red', uv ulaan 'very
red', as in Kalmuk: xav xar 'very black', whereas in Burjat, related to Mongolian, the phoneme
is /b/ instead of /v/: jab jagaan 'very pink' (Sanzeev & Todaeva 1993:139). However,
formation of the superlative by reduplication appears to be limited to certain adjectival
Among the Indo-European languages the superlative is formed with two types of
affixes: prefixes and suffixes. The most frequent were various prefixes derived from the root
*per-: per-/prae- in Latin (André 1951): praeclarus 'very famous', percallidus 'very smart';
pre-, already documented in Old Russian as pre-/pri- (Vjalkina 1995:321), still occurs with
some adjectives in Modern Russian: mudryj 'wise', premudryj 'very wise', krasnyj 'red',
prekrasnyj 'very red'; perí- in Ancient Greek: perikallé:s 'very beautiful'. In this case, the
prefix is not simply added to the positive form of adjective, but rather to the adjectival stem.
In Celtic another etymologically unclear prefix an- is attested, which already occurs in
Gallic (ande-caros 'very red'; Vendryes 1981:71), and is used in the same function: Irish an-
mhaith 'very good' (from maith), Welsh an-fawr 'very big' (from mawr; in both cases the
adjectives undergo lenition).
In some modern European languages other types of intensifying prefixed elements
occur: although their productivity is rather limited, they frequently occur in colloquial
varieties: Dutch doodmoe 'very tired', lit. 'dead-tired' (cf. Italian stanco morto), German
saublöde '(sow-silly) very silly', Italian straricco 'very rich' (stra- from Latin extra 'outside';
cf. already Latin extraordinarius 'extraordinary, beyond the limits of normality'), arcicontento
'very happy' (arci- from Ancient Greek archi- 'chief, prominent' via Latin).
The most widespread suffixes among the Indo-European languages were, however, *-
isto- and *-tato- (which bear a morphological similarity with the formatives of ordinals;
Benveniste 1948), traces of which still remain in some Indo-European branches. Their
morphosyntactic behaviour is almost identical with their comparative counterparts *-jos- and
In some languages a special prefix is added to the comparative form of the adjective or
adverb: in Hungarian, for instance, the prefix leg- is added to the comparative form: gyors-
abb 'more rapid' becomes leg-gyors-abb 'very rapid, the most rapid', the adverb gyors-an
'rapidly' leg-gyors-abb-an 'very rapidly'.
In Modern Irish, the superlative is formed by the synthetic comparative form of the
adjective preceded by the particle is: deacair 'difficult', superlative is deaicre 'most difficult',
deaicre being the comparative (see 2.3).
4. Equative
Among the graded forms, equativity is the one which exhibits the least variety. There are
basically two types, in fact: analytic and synthetic formation.
4.1. Analytic formation
In this case the parameter marker (see 1) is expressed analytically by a particle or an adverb
(12) Zürich ist so groß wie Wien.
'Zurich is as big as Vienna.'
The second is restricted to certain linguistic families and is morphological. The parameter
marker is expressed synthetically on the parameter by a special morpheme (Estonian):
(13) Minu õde on minu pikk-une.
I:GEN sister is I:GEN tall-EQT
'My sister is as tall as me.'
The standard doesn't usually occur in the case in which it occurs after the comparative,
and in some languages there is a special case marker (Ancash Quechua):
(14) Pani-i-mi qam-naw shumaq.
sister-1.SG-DIRECT.EVIDENCE you-EQT pretty
'My sister is as pretty as you.'
In some languages the parameter marker can be omitted. This is the case in Italian, for
instance, where the presence of the parameter marker is stylistically highly marked:
(15) Il mio libro è (così) bello come il tuo.
'My book is as nice as yours.'
In a few cases analytic parameter markers "are semantically more or less transparent
and mean something like 'equally, to the same degree'" (Haspelmath & Buchholz 1998:284):
(16) (a) Chinese
Ta gen ni yíyàng gao.
she with you one.manner tall
'She is as tall as you.'
(b) Seychelles Creole
I ris mem degree ki nu.
he rich same extent as we
'He is as rich as us.'
4.2. Synthetic formation
Synthetic formations are far less frequent than the analytic, i.e. syntactic ones. Among the
languages of Europe, for instance, they occur only in three linguistic groups: Celtic, Finno-
Ugric, and Kartvelian, but non-European languages also show them: Tagalog, Indonesian (in
these two languages it is a prefix; cf. (17) from Indonesian), Greenlandic Eskimo (Haspelmath
& Buchholz 1998:283f.):
(17) Ayah saya se-tinggi paman saya
father 1.SG EQT-tall uncle 1.SG
'My father is as tall as my uncle.'
In Old Irish the equative suffix was -ithir/-idir (of unclear origin; Thurneysen
1980:237f.): the choice between the two was motivated by phonology (vgl. (18a-b)): "the
former as a rule after monosyllables, the latter after polysyllables" (Thurneysen 1980:233):
(18) (a) léir 'eager'
lérithir 'as eager'
(b) erlam 'ready'
erlamidir 'as ready'
In Modern Irish the equative form is no more productive and has been replaced by the
construction [comh + the positive form of the adjective + le]: comh cliste le 'as clever as'. In
Modern Welsh, instead, the parameter marker precedes the synthetic equative form in -ed
(Middle Welsh -hed), and not the simple positive form: oer 'cold', cyn oered a 'as cold as' (not
*cyn oer a), even though this happens with some adjectives that do not have the -ed form.
5. References
Andersen, Paul K. (1983), Word order Typology and Comparative Constructions. Amsterdam: Benjamins
Andersen, Paul K. (1992), "Gradation". In: Bright, William (ed.), International Encyclopedia of Linguistics, Vol.
II. New York, Oxford: Oxford University Press, 79
André, Jacques (1951), "Les adjectifs et adverbes à valeur intensive en per- et prae-". Revue des Études Latines
29, 121-147
Austerlitz, Robert (1991), "The European Absolute Superlative, an Orphan of Grammar, of the Lexicon, and Of
History". In: Ivir, Vladimir & Kalogjera, Damir (eds.), Languages in Contact and Contrast. Berlin, New York:
Mouton de Gruyter, 1-13
Baldi, Philip & Cuzzolin, Pierluigi (2001), "Towards a New Historical Syntax of Latin". In: Moussy, Claude
(éd.), De lingua Latina novae quaestiones. Actes du Xème Colloque Internationale de Linguistique Latine, Paris-
Sèvres, 19-23 avril 1999. Louvain: Peeters, 201-225
Benveniste, Émile (1948), Noms d'agent et noms d'action en indo-européen. Paris: Maisonneuve
Bertocchi, Alessandra & Orlandini, Anna (1996), "Quelques aspects de la comparaison en latin".
Indogermanische Forschungen 101, 195-232
Bhat, D. N. S. (1994), The Adjectival Category. Amsterdam, Philadelphia: Benjamins
Booij, Geert E. (1996), "Inherent versus Contextual Inflection and the Split Morphology Hypothesis". In: Booij,
Geert & van Marle, Jaap (eds.), Yearbook of Morphology 1995. Dordrecht: Kluwer, 1-16
Borgman, Donald M. (1990), "Sanuma". In: Derbyshire, Desmond C. & Pullum, Geoffrey K. (eds.), Handbook
of Amazonian Languages, Vol. II. Berlin, New York: Mouton de Gruyter, 17-248
Chierchia, Gennaro & McConnell-Ginet, Sally (1990), Meaning and Grammar. Cambridge/MA: Massachusetts
Institute of Technology
Crookston, I. (1994), "Comparative Constructions". In: Asher, R. E. & Simpson, J. M. Y. (eds.), The
Encyclopedia of Language and Linguistics, Vol. II. Oxford etc.: Pergamon Press, 624-629
Dayley, Jon P. (1989), Tümpisa (Panamint) Shoshone Grammar. Berkeley etc.: University of California Press
Driem, George van (1987), A Grammar of Limbu. Berlin etc.: Mouton de Gruyter
Dubnova, Yelena Zinov'evna (1984), The Rwanda Language. Moscow: Nauka
Everett, Daniel L. & Kern, Barbara (1997), Wari'. London, New York: Routledge
Fortescue, Michael (1984), West Greenlandic. London etc.: Croom Helm
Galloway, Brent D. (1984), A Grammar of Upriver Halkomelem. Berkeley etc.: University of California Press
Haspelmath, Martin & Buchholz, Oda (1998), "Equative and Similative Constructions in the Languages of
Europe". In: van der Auwera, Johann (ed.), Adverbial Constructions in the Languages of Europe. Berlin, New
York: Mouton de Gruyter, 277-334
Hewitt, George B. (1995), Georgian. Amsterdam, Philadelphia: Benjamins
Hinds, John (1986), Japanese. London, New York: Routledge
Hofmann, Johann Baptist & Szantyr, Anton (1965), Lateinische Syntax und Stilistik. München: Beck
Holton, David & Mackridge, Peter & Philippaki-Warburton, Irene (1997), Greek. London: Routledge
Hudson, Grover (1997), "Amharic and Argobba". In: Hetzron, Robert (ed.), The Semitic Languages. New York:
Routledge, 457-485
Jensen, Hans (1934), "Der steigernde Vergleich und sein sprachlicher Ausdruck". Indogermanische Forschungen
52, 108-139
Klein, Henny (1998), Adverbs of Degree in Dutch and Related Languages. Amsterdam, Philadelphia: Benjamins
Koehn, Edward & Koehn, Sally (1986), "Apalai". In: Derbyshire, Desmond C. & Pullum, Geoffrey K. (eds.),
Handbook of Amazonian Languages, Vol. I. Berlin, New York: Mouton de Gruyter, 33-127
Lewis, George L. (1988), Turkish Grammar [reprinted edition with corrections]. Cambridge: Cambridge
University Press
Lyons, John (1977), Semantics. Cambridge: Cambridge University Press
Morev, Lev Nikolaevic (1988), Jazyk sèk. Moskva: Nauka
Moskalev, Aleksej Alekseevic (1971), Grammatika jazyka czuan. Moskva: Nauka
Nguyên, Dình-Hoà (1997), Vietnamese. Amsterdam, Philadelphia: Benjamins
Nikolaeva, Irina & Tolskaya, Maria (2001), A Grammar of Udihe. Berlin, New York: Mouton de Gruyter
O´Dochartaigh, Cathair (1992), "The Irish Language". In: MacAulay, Donald (ed.), The Celtic Languages.
Cambridge: Cambridge University Press, 11-99
Panagl, Oswald (1973), "Präsuppositionen und die Syntax der lateinischen Komparation". Salzburger Beiträge
zur Linguistik 1, 361-375
Pinault, Georges (1985), "Négation et comparaison védique". Bulletin de la Société de Linguistique 80.1, 103-
Pinault, Georges (1989), "Les voies de la comparaison védique". In: Valentin, Paul (ed.), La comparaison:
Linguistica Palatina, Colloquia III. Paris: Presses de l'Université de Paris-Sorbonne, 53-74
Puhvel, Jaan (1973), "Nature and Means of Comparison in Proto-Indo-European Grammar". The Journal of
Indo-European Studies 1, 145-154
Quirk, Randolph & Greenbaum, Sidney & Leech, Geoffry & Svartvik, Jan (1994), A Comprehensive Grammar
of the English Language. London: Longmann
Romero-Figueroa, Andrés (1986), "Warao Comparatives". Studies in Language 10.1, 97-108
Sanzeev, G. D. & Todaeva, B. X. (1993), "Mongolskie jazyki". In: Jazyki Azii i Afriki, V: Altajskie jazyki.
Moskva: Nauka, 98-186
Sapir, Edward (1951), "Grading: A Study in Semantics". In: Selected Writings of Edward Sapir in Language,
Culture and Personality edited by David G. Mandelbaum. Berkeley, Los Angeles: University of California
Press, 122-149 [originally appeared (1944) in Philosophy of Science 11, 93-116]
Sceglov, Jurij K. (1970), Ocerk grammatiki jazyka xausa. Moskva: Nauka
Schuh, Russell G. (1998), A Grammar of Miya. Berkeley etc.: University of California Press
Sosenskaja, Tat'jana B. (1999), "Sravnitel'nye konstrukcii". In: Kibrik, Alexander Evgen'evic (ed.), Èlementy
caxurskogo jazyka v tipologiceskom osvescenii. Moskva: Nasledie, 560-581
Stassen, Leon (1985), Comparison and Universal Grammar. Oxford: Blackwell
Suárez, Jorge A. (1983), The Mesoamerican Indian Languages. Cambridge: Cambridge University Press
Thurneysen, Rudolf (1980), A Grammar of Old Irish, trans. and revised by D. Binchy and Osborn Bergin.
Dublin: Dublin Institute for Advanced Studies
Tuite, Kenin (1997), Svan. München, Newcastle: Lincom
Vendryes, Jean (1981), Lexique étymologique de l'irlandais ancien. A. Dublin: Institute for the Advanced
Vjalkina, L. V. (1995), "Prilagatel'noe". In: Drevnerusskaja grammatika XII-XIII vv. Moskva: Nauka, 294-326
Xerberman, K. R. (1999), "Komparativnye konstrukcii v sravnenii: K voprosu ob otnosenii grammatiki k
ètimologii i jazykovoj tipologii". Voprosy jazykoznanija 2, 92-107
Pierluigi Cuzzolin, Bergamo (Italy)
Christian Lehmann, Erfurt (Germany)
... Although this difference is relevant for English, Russian and a number of other languages, the finding cannot be generalised to all languages. For one, many isolating languages and some languages with agglutinative or fusional morphology express comparison only by juxtaposing positive adjectives, as in Ôsmall boat, big boatÕ meaning that one boat is smaller than the other (Cuzzolin & Lehmann 2004). ...
This paper addresses one of the central assumptions in the analysis of scalar adjectives that their positive form is obligatorily interpreted vis-à-vis an average value. By this view, tall and taller than average are equivalent expressions. Counter to this well-established view, I argue that the two constructions are not equivalent; there are important semantic and functional differences between them. First, not all uses of the positive form are interpreted vis-à-vis an average value. Second, even in contexts where the average is relevant, it is not sufficient by itself: several standards of comparison are involved in the interpretation of scalar adjectives. Third, bare scalars and the Aer-than-average construction select different scale parts as their profile and are in this sense different construals of the gradual scale.
This paper investigates what kinds of meanings are expressed by circumfixes cross-linguistically, with an eye toward explaining this rare formal phenomenon. Based on the study of several hundred languages, it shows that circumfixes express negation relatively more often than other functions, followed by a wide array of TAME and derivational meanings and some more specific functions such as adjectival degree or the formation of ordinal numerals. While circumfixes are found in many unrelated languages and in all macro-areas, they nevertheless cluster in specific families, which suggests that it is their diachronic emergence that is rare rather than their synchronic distribution per se. The semantic contributions of circumfixes are explained by different diachronic mechanisms. Negative circumfixes are argued to ultimately follow from the negative cycle, whereas circumfixes encoding other functions are claimed to emerge from the cumulation of functionally compatible morphs, as in the case of certain aspectual and temporal markers. In addition, there is also some evidence that empty morphs may be reinterpreted as parts of circumfixes. All these processes are further constrained by the fact that one half of the developing circumfix has to be a prefix. Since prefixes are relatively rare overall, this necessary condition for circumfixation often is not met. On the other hand, prefixes express negation relatively frequently, and this helps to further explain why circumfixes have strong ties to this particular semantic domain.
This paper addresses variation and change in the realization of superlative constructions in the light of the evidence provided by Modern Greek dialectal variation as a window into the study of the organization of grammar. Dialectal data show that analyticity prevails in the realization of relative comparative constructions, while absolute ones seem to resist more persistently due to their high relevance with another morphological category, evaluative intensification. Our findings argue in favour of the strong interplay among all three processes, viewed as realizations of the conceptual category of gradation, accounted for in terms of a continuum. The proposed organization captures the strong interplay between intensification and absolute superlatives on the one hand, while relative and absolute superlative formations on the other. On a theoretical level, this account could contribute further to important issues such as the controversial status of comparison and evaluation in grammar, which may differ cross-linguistically, suggesting that a combined account of the three processes might prove more adequate.
Conference Paper
Expanding on a point made by Krifka [6, p.7-8], we show that the fact that a round number has been used significantly increases the posterior probability that that number was intended as an approximation. This increase should typically be enough to make assuming that an approximation was indeed intended a rational choice, and thereby helps explain why round numbers are often seen as simply having an approximate meaning. Generalization into non-number words is also discussed, resulting in a possible origin of (some) vagueness.
The abstract for this document is available on CSA Illumina.To view the Abstract, click the Abstract button above the document title.
Word order Typology and Comparative Constructions
  • Paul K Andersen
Andersen, Paul K. (1983), Word order Typology and Comparative Constructions. Amsterdam: Benjamins Andersen, Paul K. (1992), "Gradation". In: Bright, William (ed.), International Encyclopedia of Linguistics, Vol. II. New York, Oxford: Oxford University Press, 79
Les adjectifs et adverbes à valeur intensive en per-et prae
  • Jacques André
André, Jacques (1951), "Les adjectifs et adverbes à valeur intensive en per-et prae-". Revue des Études Latines 29, 121-147
The European Absolute Superlative, an Orphan of Grammar, of the Lexicon, and Of History
  • Robert Austerlitz
Austerlitz, Robert (1991), "The European Absolute Superlative, an Orphan of Grammar, of the Lexicon, and Of History". In: Ivir, Vladimir & Kalogjera, Damir (eds.), Languages in Contact and Contrast. Berlin, New York: Mouton de Gruyter, 1-13
Inherent versus Contextual Inflection and the Split Morphology Hypothesis
  • D N S Bhat
Bhat, D. N. S. (1994), The Adjectival Category. Amsterdam, Philadelphia: Benjamins Booij, Geert E. (1996), "Inherent versus Contextual Inflection and the Split Morphology Hypothesis". In: Booij, Geert & van Marle, Jaap (eds.), Yearbook of Morphology 1995. Dordrecht: Kluwer, 1-16
Comparative Constructions
  • I Crookston
Crookston, I. (1994), "Comparative Constructions". In: Asher, R. E. & Simpson, J. M. Y. (eds.), The Encyclopedia of Language and Linguistics, Vol. II. Oxford etc.: Pergamon Press, 624-629