LINGUIST List 12.2430

Tue Oct 2 2001

Sum: Root- vs. Affix-Markedness

Editor for this issue: Marie Klopfenstein <>


  1. Adam Ussishkin, root- vs. affix-markedness

Message 1: root- vs. affix-markedness

Date: Sun, 30 Sep 2001 07:27:55 -0700
From: Adam Ussishkin <>
Subject: root- vs. affix-markedness

The following is a summary of responses sent to me regarding a query
on the Linguist List in which I asked for examples of root-affix
asymmetries in markedness. I have done my best to provide clear and
concise paraphras es in some cases, for reasons of space. I would like
to thank the following people for taking the time to respond to my
query: Chris Beckwith, Mike Cahill, Henry Churchyard, Adrian Clynes,
Samira Farwaneh, Eloise Jelinek, Todd O'Bryan, Marc Picard, Jess
Tauber, Yishai Tobin, Pete Unseth, Suzann e Urbanczyk, Adam Werle,
Richard Wiese, Villem Wisser.

I will list below the language at issue, the markedness facts, and
references (if any) to the facts that were provided to me by the

With best wishes, and much gratitude to the respondents,

Adam Ussishkin

In Amharic, ejectives appear only in roots, and never in affixes.

In many Bantu, Niger-Congo, Central Asian agglutinating languages,
ideoph ones and expressive forms contain less marked segments (Hopper,
Paul, and Traugott, Elizabeth, 1993, Grammaticalization; Greenberg,
Joseph, 1964).

In English, inflectional affixes have only coronal consonants (except
-ing). This includes interdentals, despite the fact that they are
marked cross-linguistically.

In Chaha, laterals appear only in loanwords and affixes, never roots
(see work of Banksira, as well as Golston's 1996 Language article).

In Tibetan, syllabic affix morphemes always have simple onsets, though
complex onsets are widely attested in roots.

In Belait (an Austronesian language), initial /ny/ is found only in
pronouns (see work of Clynes, Adrian).

In English (historical development - PIE to Proto-Germanic to
Old/Middle/ Modern English) a tendency for affixes to evolve such that
they exhibit less marked segments can be seen (see work of Tobin,

In Hebrew, Hungarian, and Latin a similar trend is apparent (see work
of Tobin, Yishai).

In many native American languages, morphological affixes tend to
utilize a subset of all the consonants available within a particular
language; typically this subset comprises the less marked
phonological segments in the language (see Campbell,
Lyle. 1997. American Indian Languages. Oxford University
Press. p. 221).

In Semitic languages, affixal consonants are limited to t, k, m, n, y,
w, sh, h, and glottal stop. Crucially, marked segments such as
pharyngeals are never attested in affixes.

In German, the root-affix markedness asymmetry is clearly visible 
in both prosodic and segmental domains.

In Salish, glottalized consonants are only found in roots and lexical
suf fixes; grammatical suffixes never have glottalized consonants (see
Urbanczyk, Suzanne, 1996, Patterns of Reduplication in
Lushootseed. Ph.D. Dissertation, Umass , Amherst).

In Frisian, affixal use a limited vowel inventory; one of four sets of
suffixes are schwa-only. The same is observed in Dutch (see Visser,
Willem, The Syllable in Frisian (HIL Dissertations; 30). The Hague:
Holland Academic Graphics,

1997. ISBN: 90-5569-030-9; Booij, Geert Evert, Dutch Morphology. A
Study of Word Formation in Generative Grammar (Peter de Ridder Press
Publications on DUTCH; 2). Lisse: The Peter de Ridder Press, 1977)).

In Icelandic, erosion has taken place, such that the reflexive suffix
-sk became -st, i.e. [k] > [t]. The coronalization of this suffix
correlates with the observation that function morphemes, and
particularly inflectional morphemes, tend to contain coronal
consonants more often than consonants of other places of articulation
(see ongoing work of Werle, Adam, including a handout from HUMDRUM as
well as the full version of this paper).

In Konni, (a Gur language of Northern Ghana) the mid vowels are not
represented in any of the suffixes. The consonantal inventory of
affixes is also smaller than the full Konni set. Additionally, in a
number of languages, labiovelar consonants never appear in affixes
(see Cahill, Michael. 1999. Aspects of the Morphology and Phonology of
Konni. PhD dissertation, Ohio State Unive rsity, and Cahill, Michael,
2000, Positional Contrast and Labialvelars, OSU Work ing Papers in
Linguistics 53).

In Standard and Classical Arabic, only vowels, laryngeals, sonorant
consonants, and coronal obstruents may serve as affixal
segments. However, in some spoken dialects of Arabic an interesting
situation has arisen, whereby several affixes contain pharyngeal
consonants; these affixes have arisen from verbal stems that contain
these marked consonants. Interestingly, according to several
respondents, these 'new' affixes are changing such that the 
pharyngeals are being neutralized to laryngeal articulations in these
dialects, rendering them less marked.

* * * * * *

Adam Ussishkin
Visiting Assistant Professor
Department of Linguistics
Douglass Building, 1100 E. University Ave
University of Arizona
Tucson, AZ 85721

office: Douglass 318-B
office phone: 520-626-7121
department fax: 520-626-9014
Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue