Transcript MORPHOLOGY= The study of shapes

Morpheme:
-Abstraction
of the various types of morphs
- Smallest unit of a language that convey
some kind of information







-s in cat/ cats
-z in dog/ dogs
-es in dish/ dishes
-en in ox/ oxen
Zero morph in sheep/ sheep
Vowel change in foot/ feet
Convey the same information which also
happens in allomorphs




Free morpheme: stands alone as its own word
Bound morpheme: needs some kind of host
to attach to
He will go home tomorrow -> only free
morphemes
The oxen pulled the chart-> 2 bound
morphemes in oxen /–en/ and pulled /-ed/


Root: the smallest unit with any semantic
content
Unhappiness -> happy
Perfectly -> perfect
Stem:
- base for an inflected word form
- can consist minimally of a root but may
also a modification of the root in some ways

Example 1:
the stem horsehair is a compound of two
roots: horse + hair
This stem can me modified for plural to form:
horsehairs

Example 2:
if we add –er to the root teach we get the
stem teacher
The root and stem carry lexemic information
which is the basic semantic information of the
word.
Example:
-The lexeme of work, works, worked,
working is WORK
-The lexeme of hair, hairs is HAIR
-The lexeme of horsehair and horsehairs is
HORSEHAIR




An obligatorily bound morpheme which does
not carry any lexemic information
Affixes can be derivational or inflectional
Derivational affixes create new words like unand –ness in unhappiness
İnflectional affixes carry grammatical
information such as plural –s or the past
tense –ed and do not change the meaning of
the word



Prefix: attaches to the beginning of a host
word
Derivational prefix; un- in unhappy
İnflectional prefix: the language Logba
okpe
– inashina
o-kpe
– i-nashina
3sg-know
CM-everybody
‘He knows everybody’






Suffix: attaches to the end of the host word
Derivational suffix: -ness in happiness
İnflectional suffix: past tense –ed
İnfix: places itself inside a morpheme, usually
a root or stem
Derivational infix are used in the language
Leti where the nominalizations are derived
from the verb through the infix –niKakri ‘to cry’ -> kniakri ‘act of crying’
pali ‘to float’ -> pniali ‘act of floating’





İnflectional infix can be found in Maranao
where –i- is used to mark the past tense
Tabasan ‘slash’ -> Tiabasan ‘slashed’
Circumfix: at least two types of affixation
have to occur at the beginning and at the end
of the host at the same time.
İnflectional circumfix can be found in
German: past participle which combines of
the prefix ge- and the suffix –t
Example: lieben ‘to love’ -> geliebt ‘had
loved’





We can find derivational circumfix in
Indonesian ke- … -an which derives abstract
nouns
Example: kebebasan ‘freedom’ from the
adjective bebas ‘free’
Parafix: has two affixes which do not have to
occur at specific places like in circumfix
For example in the language Leti: the
nominalizations are derived with i- + -inatu ‘to send’ -> iniatu ‘act of sending’



Difference between clitics and affixes is that
while both are phonologically dependent on a
host, a clitic is syntactically independent from
its host while an affix is not
This means that affixes can only attach to the
kind of hosts that match their category. For
example the verbal affix –ed can only attach
to verbs. And plural affixes can only attach to
nouns and so on.
Clitics not restricted to the kind of category
they match to.





Proclitic: attaches to the beginning of the host
French pronouns may attach protoclitically:
J’attends -> 1sg=wait.PRES -> I’m waiting
Enclitic (also called postclitic) attaches at the end
of the host
Italian pronouns may attach enclitically:
E
venuto
per parl-ar=mi
3SG.is come.PFCT to talk-INF=1SG.O
‘He has come to talk to me’

Mesoclitic: attaches itself between the host
and the inflectional affixes. Very rare seen
but can be found in European Portuguese:
Pedirlheia
Pedir=lhe=ia
Ask.INF=3SG.M=1SG.COND


Endoclitic: extremely rare because it takes
place inside the root or stem
Udi and Pashto are the only two languages
which have endoclitics
Languages have been classified along a
linear scale with isolating languages on
one end, fusional languages on the other
and agglutinating languages in the middle;
adding a fourth category is introflexive.
İsolating>agglutinative> fusional> introflexive
Mandarin Turkish
Latin
Arabic
Chinese is an isolating language.
Turkish is an agglutinating language.
Linear scale merges three different parameters, fusion,
exponence and flexion.
The fourth parameter is synthesis which has to do with
how much grammatical information a word may carry.
Fusion states the degree to which morphological markers attach
to a host stem.
There are three types of fusion:
İsolating is an independent word, a marker stands alone as a free
morpheme.
Markers that are bound are concatenative. They have to attach to
a host.
Markers that involve modifying the host in some way are nonlinear.
Languages may employ any and all of the types of
fusion.
English has isolating markers ( the modal ‘must’ in He
must be home by now)
concatenative markers ( plural -s in tree (SG) versus
trees (PL))
non-linear markers
(the ablaut in sing – sang – sung).
Most languages have at least some markers that stand
in phonological isolation and thus function as
individual words. An example in English would be
the modal must, as in He must be in his office. There
are languages where all or almost all grammatical
information is conveyed though isolating markers.
Koyra Chiini:
ay
woo kaa wor o
guna
1SG.S DEM. REL. 2PL.S IPF see
‘I here whom you (PL) see.’
All grammatical information is expressed as individual
words, even the tense
of the verb (the imperfect marker o).
Concatenative means ‘chaining together’. Apart from
the fact that they are bound, they chain together in
linear strings, which means that they are
segmentable. A language with concatenative
constructions is Chichewa, where the various markers
attach linearly to the stems.
Chichewa
mlenje
mmôdzi
anabwérá
m-lenje m-môdzi a-na-bwérá
ndí
míkôndo
ndí mí-kôndo
I-hunter ı.sm-one ı.sm-past-come with ıv-spears
‘One hunter came with spears.’
The grammatical markers for noun class (ı m/a and ıv mi) and
past tense (na) are bound and are relatively straightforward to
segment into morphemes.
Non-linear markers involve some kind of modification to
the host stem and are not straightforward to segment into
chains of morphemes. Languages modify their stems non-
linearly. A root consists only of a set of consonants and
grammatical information is conveyed through insertion of a
pattern of vowels, commonly termed the “root-and-pattern”
but it is also termed ablaut. Neither the root nor the vowel
pattern can function on its own. Modern Hebrew is a language
with such a pattern;
g-d-r ‘enclose’
past:
a-a
(CaCaC):
gadar
‘enclosed’
present:
o-e
future: yi-Ø-o
(CoCeC):
goder
(yiCCoC):
yigdor ‘will enclose’
imperative: Ø-o (CCoC):
infinitive: li-Ø-o (liCCoC):
enclose’
gdor
ligdor
‘encloses’
‘enclose!’
‘to
Another example of ablaut (also called gradation or
vowel gradation) is found in the strong verbs in Germanic
languages, where inflection is marked through changes in
the root vowel quality, as in English sing – sang – sung
(present – past – past participle).
Suprasegmentals (or prosodic formatives), involving
tone, stress and length, are another type of non-linear
morphological
processes.
Tone
is
a
well-known
morphological strategy, An example of a language with
grammatical tone is Lango.
a- àpônnê
1SG.hide.PFV.MID
1SG.hide.PROG.MID
‘I hide myself.’
myself.’
b-
ápònnê
‘I am hiding
Replacement or substitution is a regular marker and
replaces a part of the stem.
Another type of replacement is suppletion, where a
root or stem is replaced by a root or stem of a different
etymological origin.
A rare type of non-linear process is subtraction, where
the grammatical information lies in taking out an
element of the stem.
Reduplication falls consomewhere in between
concatenation and non – linear process.
It involves copying a set amount of phonological
material from a base form (root or stem) and fusing
it with that base to form a stem onto which other
morphemes may then be added.
It’s less linear than concatenative morphemes in that
the form of the reduplicant (the repeated element) is
dependent on the form of the base, since it is a part
of the base that is being repeated.




Reduplication can be either full or partial, and while
the reduplicant usually attaches immediately to the
root it has its shape from, there are also languages
with socalled discontinuous reduplication, where
other morphological material may appear between
the reduplicant and the base.
Also, reduplication can be simple or complex.
In simple reduplication merely repeats a given
amount of material from the base.
Complex reduplication involves taking material from
the base and partly altering it.

Full reduplication involves copying the whole base.
Most languages allow both full and partial
reduplication.

In Rubino’s database shows only 35 languages
which are allow full reduplication. Here is an
example of a language with full simple reduplication
is Erromangan, where reduplication indicates
intensification.
Erromangan (Austronesian (Oceanic): Vanuatu
/unmeh/ “early” >> /unmehunmeh/ “very early”
/ilar/ “shine” >> /ilarilar/ “shine brightly”
(Crowley 1998:34)


An example of a full complex reduplication can be
found in Persian, where the reduplicated form
changes the initial consonant to either /m/ or /p/ of
copied element. The reduplicated form takes a
meaning of what we might call “scattered generality”,
most closely equivalent to English ‘and so forth’.
Persian ( Indo – European (Iranian): Iran)
bâlâ “ above” >> bâlâmala “somewhere above”
mive “fruit” >> mivepive “fruit and so on”
(Ghaniabadi et al. 2006:3)


Partial reduplication involves copying only a set part
of the base and may involve a number of different
forms. It can be a set of phonemes (C,CV,CVCV, and so
on) , a set of syllabes or a set of morae (the minimal
unit of metrical weight) that is copied. (Rubino, 2011)


In Thao the instrumental is expressed by Careduplication, which means that the first consonant
of the base is copied and -a- is added (also called
duplifix, Haspelmath 2002:24):
Thao (Austronesian (Paiwanic): Taiwan)
cput “to filter” >> cacput “sieve”
>> c - a - cput
An example of a partial complex reduplication can be
found in Nakanai;
velo “bubbling” >> velelo “bubbling forth”
ve - le - lo

Automatic reduplication is when an affix obligatorily
triggers reduplication but the reduplication itself does
not add any meaning to the construction. An example
of an automatic reduplication can be foun also in
Tagalog;
Tagalog (Austronesian (Meso – Philippine): Philippines:
wilih “interested” >> kawilihwilih “interested”
ka – wilih – wilih
(French 1998:50)

As it is mentioned above, the reduplicant might
be seperated from the base by some particle. An
example of such a discontinuous reduplication can be
found in the Manila Bay Creoles, which is a cover
term for Ternateno, Caviteno, and Ermiteno, where
the linker - ng - sits between the reduplicant and the
base.
Manila Bay Creoles (Creole (Spanish – lexified): Philippines)
Bunita “beautiful” >> bunitangbunita “very beautiful”
bunita – ng – bunita
(Grant 2003:205)

All these examples implies that pidgings and
creoles do not seem to behave differently from non –
creole languages in terms of employing the
morphological process of reduplication.






In Turkish, the process of emphatic
reduplication, the purpose of which is to
accentuate the quality of
an adjective, involves the copying of the initial
(C)V of the base and then prefixing it, along with
an
additional affixal consonant from the set /p, s,
m, r/, to the base, as seen in (1). In some cases,
the
emphatic (C)VC prefix is also followed by –A, –Il,
or –Am, as seen in (2). Cases such as those in (2)
are
considered idiosyncratic and are not the result of
a productive phonological process (Göksel and
Kerslake
2005).
(1) güzel ‘pretty’
uzun ‘long’
katı ‘hard’
siyah ‘black’
temiz ‘clean’
güpgüzel ‘very pretty’
upuzun ‘very long’
kaskatı ‘hard as a rock’
simsiyah ‘pitch black’
tertemiz ‘clean as a pin’
(2) gündüz ‘daytime/by day’
broad daylight’
yalnız ‘alone’
çıplak ‘naked’
parka ‘piece’
to
pieces’
güpegündüz
‘in
yapayalnız ‘all alone’
çırılçıplak ‘stark naked’
paramparça ‘torn to
shreds/smashed
Doubling occurs in two ways: simple doubling and
doubling in lexical formations. In simple doubling,
the word is repeated. Depending on the syntactic
category of the targeted lexeme, it can produces
adverbials, adjectivals and measue terms (Göksel &
Kerslake 2005).
 tek tek
zaman zaman
one DUP
time DUP
“one by one”
“time to time”

Some additional morphemes, such as plural
suffix and the question particle, are attached to
the sister conctituents or one of the
constituents undergoes phonetic changes for
doubling in lexical formations
 güzel-ler
güzel-i bir kız
beautiful-PLU
beautiful-POSS a girl
‘a very beautiful girl
 güzel mi
güzel bir kız
beautiful QP beautiful a girl
‘a very beautiful girl’
 ufak tefek bir kutu
little fi(little) a box
‘a tiny box’

Examples:
 With Synoynms; güçlü kuvvetli, ses seda, sağlık
sıhhat, evirmek çevirmek etc.
 With nearly the same meanings; eş dost, doğru
dürüst, ağrı sızı, sağ salim etc.
 With antonyms; iyi kötü, aşağı, yukarı, irili
ufaklı, acı tatlı etc.
 With meaningless words; abuk subuk, abur
cubur, eciş bücüş, apar topar etc.
 With Onomatopoeia words; tıkır tıkır, şırıl şırıl,
horul horul, vızır vızır etc.

Languages also differ as to how many grammatical categories
may be expressed by one and the same morpheme. Seperative
morphemes(or monoexponential) morphemes encode only one
single category, cumulative (polyexponential, also called
portmanteau) morphemes encode several things at the same time.
This parameter may interact with fusion, so that we get six logical
logical combinations: isolating, concatenative, and non-linear
seperative markers plus isolating, concatenative, and non-linear
cumultative markers.
Here is a list for languages with examples of each of the six
logical types of processes.
Kasong: isolating seperative
Meithei: concatenative seperative
Dinka: non-linear seperative
Wari: isolating cumulative
Spanish: concatenative cumulative
Hebrew: non-linear cumulative
For Kasong language, each of the markers is a free morpheme.
They are isolating , and each of them conveys only one piece of
information, the markers are seperative:
Kasong
nak kamlaŋ loŋ ce:w prǐ
3.SG PROG. FUT. go forest
‘s/he will be going to the forest.’
Meithei(Sino-Tibean (Kuki-Chin): India) offers an example of
concatenative seperative markers. The markers fuse
concatenatively with a host stem; they are linearly segmentable
and each of the segments in that each conveys only
one piece of information.
Meithei
ǝynǝ thǝŋ
ǝy-nǝ thǝn
ǝmǝnǝ hǝydu kháy
ǝ-mǝ-nǝ
hǝy-tu kháy-i
1.SG-CNTR
knife
ATT-one-INST
‘ I cut the fruit with a knife.’
fruit-DDET cut-NHYP
(Chelliah 1997:128)
Dinka language has non-linear seperative process, where the
absolute and locative cases are distinguished only through
phonological length. The marker conveys only the information of
case, and is as such seperative, but it is not possible to segment
from the host word, and is therefore non-linear.
Dinka
tôoc
tôooc
‘swampy.area.ABSOLUTIVE
---‘swampy.area.LOCATIVE’
(ANDERSEN 2002: 13)
Wari is an example a language with isolating markers, that is the
morpheme form seperate words. However, they are cumulative in
that they contain more than one piece of grammatical information,
and this information is not possible to segment into smaller units.
Wari
ma’
co
that.PROX.HEARER
‘ Who is speaking? ’
tomi
INFL.M/FRP/P
na
speak
3SG.RP/P.VIC
(Everett 1998: 692)
Spanish also makes use of cumulative markers that fuse
concatenatively onto the stem, which gives us a concatenative
cumulative morphological process.
Spanish
habl-ó
speak-3sg.past.ind.pfv
‘He spoke.’
(source: personal knowledge)
In Hebrew has a non-linear cumulative process. It means a
similar amount of information is expressed through only one
single process, but the process involves modifying the root
itself and is thus non-linear.
Modern Hebrew
g-d-r ‘enclose’
future active indicative: yigdor ‘will enclose’
future passive indicative: yigader ‘will be enclosed’
(Glinert 1989: 471)
the way the stem is modified conveys more than one piece of
information: the tense, the voice, and the mood.
However, this grammatical information is not segmentable: if you
want to change any of the grammatical information, for instance
from active voice to passive you have to modify the root to an
entirely different stem.
FLEXIVITY
Languages also differ in how much allomorphy they have, termed
flexitivity in Bickel & Nichols (2007). The Indo-European
declension and conjugation classes are examples of flexitivity.
That is where a set of infectional affixes are chosen depending
on which class the noun or verb belongs to.
On the other hand, a given grammatical marker is always the
same. It does not vary according to classes of verbs or nouns, it
is nonflexive. If a language has five different ways of marking the
(nominative) plural, with -e, -er, (e)n -s, or –Ø, depending on
which class the noun belongs to, we have an instance of
flexitivity. It exhibited in German. If the plural is always marked
the same way, as is the case with Pichi dέn (Yakpo 2009), we have
an instance of nonflexitivity.
This is a third and seperate parameter from fusion and
exponence and may interact with them in various ways. We have
four logical combinations with the languages exemplifying types
included.
Cumulative
English
Separative
Flexive
German
Nonflexive
Hawai‘i Creole
Warlpiri
Pichi
German is an example of flexive cumulative morphemes. Because ,
the choice of which allomorph to take depends on which declension
class the noun belongs to flexitivity and the markers express both
number and case cumulative.
An example of a nonflexive cumulative marker is the Hawai’i Creole
English. For
example; wεn which expresses both tense (past) and aspect
(perfective) at the
same time. It is cumulative.
The plural marker in Pichi is an example of an nonflexive seperative
marker
because it is invariant as the plural marker nonflexive and it means
only plural and nothing else (seperative).
An example of a flexive separative marker can be found in Warlpiri
where the ergative case is marked either with -ngku or with –rlu. It is
flexive in that there are two alternative ways of marking ergative
case, and it is separative in that it means only one thing (ergative).
Likewise, flexitivity interacts with fusion. %e German plural
marking mentioned above is both flexive and concatenative; this
is, in fact, the most common combination. Flexive nonlinear
strategies are common in Semitic languages;
we have seen that Hebrew expresses tense, mood and voice
through a set of vocalisms. Flexive isolating markers are very rare
but can be found in Sierra Otomí, where person and tense is
marked by a free morpheme which looks different depending on
what conjugation class the verb belongs to:
Sierra Otomí
1sg.pres
verb
conjugationclass
dí
petsi ‘I keep (it)’
I
dín
tófo ‘I say (it)’
II
dídí
hóqui ‘I -x (it)’
III
dídím
pepfi ‘I work’
IV
The Pichi plural marking mentioned above is an example of a
nonflexive isolating marker. This is is pretty typical: “[n]onflexive
formatives are often isolating; and the most common type of
isolating formative is nonflexive.
(Bickel &Nichols 2007:187)
Turkish is an example of a language where the plural marker -lar is
nonflexive concatenative – also a very common strategy – as it
attaches to a host but is segmentable, and is invariable, i.e. is used
for all nouns
(Kornfilt 2003: 265)
An example of a nonflexive non-linear marker is the perfective
marker in Kisi. It invariably expressed through a LH tone
(Childs 1995: 173).
Here is a table summarizing for the six logical combinations with the
languages exemplifying each type included.
Flexive
Nonflexive
Isolating
Sierra Otomí
Pichi
Concatenative
German
Turkish
Non-linear
Hebrew
Kisi
What we have seen is that languages employ different strategies.
And that these strategies themselves fall along three separate
parameters that all interact with each other.
Another parameter is that of Synthesis, which can
thought of as a scale indicating how much accumulated
information a word can hold, as opposed to the
parameters given above, which, again very simplified,
basically denote what kinds of morphemes languages
tend to have and how they combine. But bear in mind
that I am simplifying matters considerably by merging
the concepts of phonological word and grammatical
word.
There are three basic types of synthesis, which can be
pictured as standing in a linear arrangement to each
other.
ANALYTIC > SYNTHETIC > POLYSYNTHETIC
ANALYTIC words do not take any affixation to their
lexical roots or stems. An analytic way of marking
tense, for example, is found in the English future, as in
He will walk home.
SYNTHETIC words allow affixation. An example of
synthetic tense in English is the past, expressed through
the –ed affixation, as in He walked home.
English typically does not take a high
amount of affixation. For instance, while the
grammatical coding of comparative for
adjectives tends to be done synthetically if
the stem is rather short, an analytic
construction is favourred if the stem is rather
long.
The Chichewa Example also shows
instances of synthetic words, where several
pieces of grammatical information are
attached to the lexical root or stem. But a
synthetic word can also end up being very
long. A spectacular case of synthesis can be
found in Turkish.
TURKISH ( Altaic ( Turkic ): Turkey )
tanıştırılamadıklarındandır
tan-ış-tır-ıl-a-ma-dık-lar-ın-dan-dır
“it is because they cannot be introduced to each other.”
The crucial difference between synthetic and
polysynthetic words is that the latter involve
more than ne lexeme. While the Turkish
example is very long and involves a great deal
of segments, there is only one lexeme, tan
‘know’. Polysynthetic words, however, may
contain more than one lexeme. Alutor is an
example of a language with polysythetic
words.
ALUTOR (Chulotko- Kamchatkan ( Northern
Chukotko- Kamchatkan): Russia)
gəmmə
gəmmə
tk-ən
takkannalgənkuwwatavətkən
t-akka-n-nalgə-n-kuww-at-avə-
‘I am making a son dry a skin/skins.’
The Turkish word tanıştırılamadıklarındandır
in as long as the Alutor word
takkannalgənkuwwatavətkən but the Turkish
word is synthetic while the Alutor word is
polysynthetic. This is because the Alutor word
contains three different lexemes, akka ‘son’,
nalgə ‘skin’ and kuww ‘dry’. Although
polysynthetic words tend to be long, they do
not necessarily have to be as following.
Mamaindê ( Nambikuaran
(Nambikuaran):Brazil)
Jukhoʔth ɪ̈ntu
Ju-khoʔ-th ɪ̈n-tu
‘village hanging on the edge’
The mamainde word is shorter than the
Turkish word but is still a case of
polysynthesis, since it contains two lexemes
ju ‘edge and khoʔ ‘hang’.
Sign languages, just like spoken languages,
have minimal meaningful units, i.e.
morphemes, and instances where units may
alternate, i.e. allomorphy. Morphemes may
either free or bound. In other words signed
languages are as linguistically complex as
spoken languages. However, due to the fact
that sign languages make use of an entirely
different mode of communication, visual
instead of audio, morphology in sign
language tends to be less concatenative than
in spoken languages.
Compounding, which is also sequential in
nature, is very common in sign languages. An
example of a compound is the ASL sign for
faint which consists of the signs MIND+DROP.
An example of a derivation is the ISL negative
suffix, which,similar to the English –less,
derives adjectives, for instance shameless in
the construction SHAME + neg. This negative
suffix has two allomorphs, signed either with
one hand or two, depending on the host it
attaches to.
Examples of prefixes are the ISL.’sense’
prefixes: to denote that something has to do
with perception ( seeing/hearing/smelling) a
reduced and bound EYE-SHARP ‘to discern by
seeing’. Examples of cliticized forms occur in
Turkish Sign Language ‘TİD’ and DGS. In TİD
the negator NOT may attach itself to the
preceding sign and form part of a
phonological unit with that host: it
(en)cliticizes.
Non-linear morphological processes are
very common in sign languages. For example,
verbs are very often modified non-linearly for
agreement with the subject and object or for
aspect. What is non-linear about much of
sign language and morphology is that the
base of the sign, the stem, is modified as to
its rhythm, path or direction to indicate the
relevant grammatical information.
It seems as if sign languages universally
make use of what has been termed
classifiers. They modify verbs and typically
decode the shape of objects, the handling of
an object and the movement and location of
referents. With classifiers, “ the handshape of
one or both hands represents a particular
type of referent, while the location,
arrangement and movement of the hand
expresses something about referrent”.
These classifiers are organized
paradigmatically. An example of a complex
sign using clasifiers would be expressing the
sentence The car hits a tree ( and gets
wrekced) in ASL. Here non-dominant hand is
configured for the clasifier “tree” while the
dominant hand is configured for “vehicle”,
signs “move” and adds the configurations for
“wrecked” at the end of the motion.
There are two major types of classifiers, entity
classifiers and handling classifiers. Sign
languages vary in the amount of classifiers
they have. For example, NGT has 17 classifier
hand-shapes while Indo-Pakistani Sign
Language only has two, “legs” and “person”.
Many sign languages make use of
reduplication to express the general concept
of “more of the sasme”, similarly as in spoken
languages.
Sign reduplication is done by having the
sign make an arch and thereby repeating the
location-movement-location pattern in one
fluid motion. A reduplicated verb will typically
indicate a longer duration of the event
(durative), or that is occurs
habitually(habitual), or that it occurs
repeatedly(iterative). A reduplicated noun
typically indicated plurality.
Both spoken and signed languages make
use of morphemes. These can be either
bound or free. The core of a lexeme is a root
or a stem, the difference between the two
being that the root is not further analysable
into any smaller parts, while a stem may
consist of a root plus something else. Affixes
are bound morphemes that do not carry any
lexemic information and that are syntactically
dependent on what kind of host they may
attach to.
Clitics are also bound morphemes, but while
they are phonologically dependent on a host,
they are not syntactically dependent on what
they may attach to. Both affixes and clitics
can attach at different places on their hosts.
Fusion indicates how tightly morphemes
attach to each other. Reduplication is a kind
of fusion. Exponence indicates how much
information each morpheme conveys.
Flexion denotes how much allomorphy a language has. A
seperate, fourth, parameter is that of synthesis, which
denotes how much information, both grammatical and
lexemic, a word may carry. Sign languages are as
morphologically complex as spoken languages, but due to
their difference in modality- spoken languages being
dependent on the sequential nature of sound while signed
languages have at their disposal the simultaneity of the
visual medium- spoken languages are predominantly linear
in their morphological processes while signed languages are
predominantly non-linear.