Roots unrooted
Pavel Caha
The morphologist’s view: Roots vs.
affixes
1/28
Roots: the morphologist’s view
Aronoff (1994): A root is what is left when all morphological structure
has been wrung out of a form.
(1) Czech conjugation
sg. pl.
1st prosím prosíme
2nd prosíš prosíte
3rd prosí prosí
2/28
Roots: the morphologist’s view
Aronoff (1994): A root is what is left when all morphological structure
has been wrung out of a form.
(1) Czech conjugation
sg. pl.
1st prosí-m prosí-me
2nd prosí-š prosí-te
3rd prosí prosí
2/28
Roots: the morphologist’s view
Aronoff (1994): A root is what is left when all morphological structure
has been wrung out of a form.
(1) Czech conjugation
sg. pl.
1st prosí-m prosí-me
2nd prosí-š prosí-te
3rd prosí prosí
past.masc prosil prosili
inf prosit prosit
2/28
Roots: the morphologist’s view
Aronoff (1994): A root is what is left when all morphological structure
has been wrung out of a form.
(1) Czech conjugation
sg. pl.
1st pros-í-m pros-í-me
2nd pros-í-š pros-í-te
3rd pros-í pros-í
past.masc pros-i-l pros-i-l-i
inf pros-i-t pros-i-t
2/28
Roots: the morphologist’s view
Aronoff (1994): A root is what is left when all morphological structure
has been wrung out of a form.
(1) Czech conjugation
sg. pl.
1st pros-í-m pros-í-me
2nd pros-í-š pros-í-te
3rd pros-í pros-í
past.masc pros-i-l pros-i-l-i
inf pros-i-t pros-i-t
imp pros pros-te
2/28
Roots: the morphologist’s view
Aronoff (1994): A root is what is left when all morphological structure
has been wrung out of a form.
(1) Czech conjugation
sg. pl.
1st pros-í-m pros-í-me
2nd pros-í-š pros-í-te
3rd pros-í pros-í
past.masc pros-i-l pros-i-l-i
inf pros-i-t pros-i-t
imp pros pros-te
2/28
The M-Root
In a lot of words, one can distinguish the root and the affixes
(2) Czech demonstratives
fem. neut. masc.
nom ta to ten
acc tu to toho
gen té toho toho
loc té to toho
dat té to toho
ins tou tím tím
3/28
The M-Root
In a lot of words, one can distinguish the root and the affixes
(2) Czech demonstratives
fem. neut. masc.
nom t-a t-o t-en
acc t-u t-o t-oho
gen t-é t-oho t-oho
loc t-é t-o t-oho
dat t-é t-o t-oho
ins t-ou t-ím t-ím
3/28
The M-Root
In a lot of words, one can distinguish the root and the affixes
(2) Czech demonstratives
fem. neut. masc. wh, anim.
nom t-a t-o t-en
acc t-u t-o t-oho k-oho
gen t-é t-oho t-oho k-oho
loc t-é t-o t-oho k-om
dat t-é t-o t-oho k-omu
ins t-ou t-ím t-ím
3/28
The syntactician’s view: Lexical
categories vs. Functional catgories
4/28
Lexical categories
• In syntax, people used to have a related distinction, namely that
between LEXICAL CATEGORIES and FUNCTIONAL CATEGORIES.
5/28
Lexical categories
• In syntax, people used to have a related distinction, namely that
between LEXICAL CATEGORIES and FUNCTIONAL CATEGORIES.
(3) Syntactic structures
(Chomsky (1957))
a. NP → T + N
b. T → the
c. N → man, ball, ...
(4) NP
T
the
N
man
5/28
Lexical categories
• In syntax, people used to have a related distinction, namely that
between LEXICAL CATEGORIES and FUNCTIONAL CATEGORIES.
(3) Syntactic structures
(Chomsky (1957))
a. NP → T + N
b. T → the
c. N → man, ball, ...
(4) NP
T
the
N
man
• One distinction is obvious from the notation: open vs. closed
class items.
5/28
The DP hypothesis
(5) Abney (1987)
DP
D
the
N
book
6/28
Affixes can be functional heads
(6) Danish
a. en
a
bog
book
b. bog-en
book-def
‘the book’
c. den
the
gamle
old
bog
book
7/28
Affixes can be functional heads
(6) Danish
a. en
a
bog
book
b. bog-en
book-def
‘the book’
c. den
the
gamle
old
bog
book
(7) DP
D
den
NP
AP
gamle
N
bog
(8) DP
N
bog
D
en 7/28
Affixes can be functional heads
(6) Danish
a. en
a
bog
book
b. bog-en
book-def
‘the book’
c. den
the
gamle
old
bog
book
(7) DP
D
den
NP
AP
gamle
N
bog
(8) DP
N
bog
D
en
Lexical categories are M-roots.
7/28
Affixes can be functional heads
(6) Danish
a. en
a
bog
book
b. bog-en
book-def
‘the book’
c. den
the
gamle
old
bog
book
(7) DP
D
d-en
NP
AP
gamle
N
bog
(8) DP
N
bog
D
en
Lexical categories are M-roots.
Functional categories may be
affixal (but do not have to be).
7/28
How lexical categories became
empty
8/28
Plural
(9) the books
9/28
Plural
(9) the books
(10) Chomsky (1957)
a. NP → NPsing
b. NP → NPpl
c. NPsing → T + N + Ø
d. NPpl → T + N + S
(11) NP
T
the
N
book
S
s
9/28
Plural
(9) the books
(10) Chomsky (1957)
a. NP → NPsing
b. NP → NPpl
c. NPsing → T + N + Ø
d. NPpl → T + N + S
(11) NP
T
the
N
book
S
s
(12) DP
D
the
NumP
N
book
Num
-s 9/28
Gender
(13) Spanish
a. l-a-s
the-fem-pl
muchach-a-s
child-fem-pl
‘the girls’
b. l-o-s
the-masc-pl
muchach-o-s
child-masc-pl
‘the boys’
10/28
Gender
(13) Spanish
a. l-a-s
the-fem-pl
muchach-a-s
child-fem-pl
‘the girls’
b. l-o-s
the-masc-pl
muchach-o-s
child-masc-pl
‘the boys’
(14) DP
D
las
NumP
GenderP
N
muchach
Gender
-a
Num
-s
10/28
Portmanteau
(15) Luganda
a. omu-ntu ‘person’ (class 1)
b. aba-ntu ‘people’ (class 2)
c. eki-ntu ‘thing’ (class 7)
d. ebi-ntu ‘things’ (class 8)
e. awa-ntu ‘place’ (class 16)
11/28
Portmanteau
(15) Luganda
a. omu-ntu ‘person’ (class 1)
b. aba-ntu ‘people’ (class 2)
c. eki-ntu ‘thing’ (class 7)
d. ebi-ntu ‘things’ (class 8)
e. awa-ntu ‘place’ (class 16)
(16) a. ss-a-yas-izza
neg.1-past-break-perf
ki-kopo
7-cup
‘I didn’t break any cup.’
b. ss-a-ky-as-izza
neg.1sg-past-7oc-break-perf
e-ki-kopo
7-7-cup
‘I didn’t break the cup.’
11/28
Portmanteau
(15) Luganda
a. o-mu-ntu ‘person’ (class 1)
b. a-ba-ntu ‘people’ (class 2)
c. e-ki-ntu ‘thing’ (class 7)
d. e-bi-ntu ‘things’ (class 8)
e. a-wa-ntu ‘place’ (class 16)
(16) a. ss-a-yas-izza
neg.1-past-break-perf
ki-kopo
7-cup
‘I didn’t break any cup.’
b. ss-a-ky-as-izza
neg.1sg-past-7oc-break-perf
e-ki-kopo
7-7-cup
‘I didn’t break the cup.’
11/28
Bantu II
DP
D
a
NumP
Num
Num Gen
N
ntu
12/28
Bantu II
DP
D
a
NumP
Num⇒ ba
Num Gen
N
ntu
12/28
Once you factor functional structure away, there is
nothing left
• Borer: Thus far, the investigation of e.g. the table or walked the
dog proceeded from the assumption that formal properties of
such expressions can be fully accommodated without availing
ourselves, at any point, of information that is uniquely
connected to table, walk and dog respectively. Rather, both
syntax and the crucial aspects of the semantics can be
computed on the basis of functors and the semantic formulas
which such functors denote.
13/28
Once you factor functional structure away, there is
nothing left
• Borer: Thus far, the investigation of e.g. the table or walked the
dog proceeded from the assumption that formal properties of
such expressions can be fully accommodated without availing
ourselves, at any point, of information that is uniquely
connected to table, walk and dog respectively. Rather, both
syntax and the crucial aspects of the semantics can be
computed on the basis of functors and the semantic formulas
which such functors denote.
• A lot of people in Generative Grammar now entertain this
position.
13/28
The birth of S-roots
14/28
A-categorial categories
dance: N / V
dog: N / V (to cause trouble for someone over a long period of time)
...
15/28
A-categorial categories
dance: N / V
dog: N / V (to cause trouble for someone over a long period of time)
...
Harley and Noyer (1999): ... different “parts of speech” can be
defined as a single l-morpheme, or Root (to adopt the terminology of
Pesetsky 1995), in certain local relations with category-defining
f-morphemes. For example, a noun or a nominalization is a Root
whose nearest c-commanding f-morpheme (or licenser) is a
Determiner ...
15/28
A-categorial categories
dance: N / V
dog: N / V (to cause trouble for someone over a long period of time)
...
Harley and Noyer (1999): ... different “parts of speech” can be
defined as a single l-morpheme, or Root (to adopt the terminology of
Pesetsky 1995), in certain local relations with category-defining
f-morphemes. For example, a noun or a nominalization is a Root
whose nearest c-commanding f-morpheme (or licenser) is a
Determiner ...
(17) DP
D
the dog 15/28
A-categorial categories
dance: N / V
dog: N / V (to cause trouble for someone over a long period of time)
...
Harley and Noyer (1999): ... different “parts of speech” can be
defined as a single l-morpheme, or Root (to adopt the terminology of
Pesetsky 1995), in certain local relations with category-defining
f-morphemes. For example, a noun or a nominalization is a Root
whose nearest c-commanding f-morpheme (or licenser) is a
Determiner ...
(17) DP
D
the dog
(18) TP
dog
T
-ed 15/28
“A novel syntactic term”
Borer, Structuring Sense III, 347: by virtue of being syntactic objects
without a category, roots represent a novel syntactic term.
(19) a. S-roots: Whatever is left when functional categories are
“wrung out” of a form.
b. M-roots: Whatever is left when all morphological
structure has been wrung out of a form.
16/28
“A novel syntactic term”
Borer, Structuring Sense III, 347: by virtue of being syntactic objects
without a category, roots represent a novel syntactic term.
(19) a. S-roots: Whatever is left when functional categories are
“wrung out” of a form.
b. M-roots: Whatever is left when all morphological
structure has been wrung out of a form.
(20) DP
D
d-en
NP
AP
gamle
N
bog
16/28
My position
• There are no S-roots — There are no a-categorial root nodes in
syntax, no novel objects in the Borer/DM sense
17/28
My position
• There are no S-roots — There are no a-categorial root nodes in
syntax, no novel objects in the Borer/DM sense
• There are no lexical categories
17/28
My position
• There are no S-roots — There are no a-categorial root nodes in
syntax, no novel objects in the Borer/DM sense
• There are no lexical categories
• Functional categories all the way down
17/28
My position
• There are no S-roots — There are no a-categorial root nodes in
syntax, no novel objects in the Borer/DM sense
• There are no lexical categories
• Functional categories all the way down
• M-roots are the pronunciation of functional categories
17/28
My position
• There are no S-roots — There are no a-categorial root nodes in
syntax, no novel objects in the Borer/DM sense
• There are no lexical categories
• Functional categories all the way down
• M-roots are the pronunciation of functional categories
17/28
Predecessors
(21) Ramchand (2008)
initP
initiation ProcP
process ResP
Result
18/28
Predecessors
(21) Ramchand (2008)
initP
initiation ProcP
process ResP
Result
• enter = init+proc+res
• walk = init+proc
• melt = proc
18/28
Nanosyntax and adjectives
19/28
The framework
• The basic building blocks of language are very small (smaller
than morphemes)
• Phrasal spell-out (and the Superset Principle)
20/28
The framework
• The basic building blocks of language are very small (smaller
than morphemes)
• Phrasal spell-out (and the Superset Principle)
(22) F5
F5 F4
F4 F3
F3 F2
F2 F1
20/28
The framework
• The basic building blocks of language are very small (smaller
than morphemes)
• Phrasal spell-out (and the Superset Principle)
(22) F5 ⇒ phon
F5 F4
F4 F3
F3 F2
F2 F1
20/28
The framework
• The basic building blocks of language are very small (smaller
than morphemes)
• Phrasal spell-out (and the Superset Principle)
(22) F5 ⇒ phon
F5 F4
F4 F3
F3 F2
F2 F1
(23) F5
F5 F4
F4 F3
F3 F2
F2 F1
⇔ phon
20/28
Two classes of adjectives
(24) a. gradable: tall, rich, fast, warm,...
b. non-gradable: nuclear, communal, baroque, ...
21/28
Two classes of adjectives
(24) a. gradable: tall, rich, fast, warm,...
b. non-gradable: nuclear, communal, baroque, ...
(25) a. a very/extremely warm weather
b. *a very/extremely nuclear family
21/28
Two classes of adjectives
(24) a. gradable: tall, rich, fast, warm,...
b. non-gradable: nuclear, communal, baroque, ...
(25) a. a very/extremely warm weather
b. *a very/extremely nuclear family
(26) a. a very American movie
b. an un-American movie
21/28
Two classes of adjectives
(24) a. gradable: tall, rich, fast, warm,...
b. non-gradable: nuclear, communal, baroque, ...
(25) a. a very/extremely warm weather
b. *a very/extremely nuclear family
(26) a. a very American movie
b. an un-American movie
(27) a. This stipend is for (*very) American scientists
b. This stipend is for non-American scientists
21/28
Two classes of adjectives
(24) a. gradable: tall, rich, fast, warm,...
b. non-gradable: nuclear, communal, baroque, ...
(25) a. a very/extremely warm weather
b. *a very/extremely nuclear family
(26) a. a very American movie
b. an un-American movie
(27) a. This stipend is for (*very) American scientists
b. This stipend is for non-American scientists
(NB: I came up with the examples myself on the basis of descriptions and the negative examples seem not to be correct, as
pointed out by Jeff)
(28) Gradable adjectives are based on scales, non-gradable adjectives have no such scale. A non-gradable adjective
can be turned into a gradable adjective by associating a scale to it.
21/28
Two classes of gradable adjectives
(29) Gradable adjectives form pairs belonging to an identical scale
a. happy — sad
b. friendly — hostile
c. healthy — sick
22/28
Two classes of gradable adjectives
(29) Gradable adjectives form pairs belonging to an identical scale
a. happy — sad
b. friendly — hostile
c. healthy — sick
(30) Positive vs. negative adjectives
22/28
Two classes of gradable adjectives
(29) Gradable adjectives form pairs belonging to an identical scale
a. happy — sad
b. friendly — hostile
c. healthy — sick
(30) Positive vs. negative adjectives
a. unhappy — *unsad
b. unfriendly — *unhostile
c. unhealthy — *unsick
22/28
The three classes of adjectives
Ongoing work by Guido Vanden Wyngaerd and Karen de Clercq
(31) a. sad
NegP
Neg ScaleP
Scale PropertyP
...
b. happy
ScaleP
Scale PropertyP
...
c. nuclear
PropertyP
Property
23/28
No difference between roots/affixes
(32) Deriving adjectives
a. positive adjectives: (un-)event-ful, (un-)faith-ful,
(un-)help-ful, (un-)law-ful, (un-)success-ful
b. negative adjectives: (*un-)use-less, (*un-)breath-less,
(*un-)sense-less, (*un-)merci-less, (*un-)cheer-less
24/28
In tree structure
(33) a. -less
NegP
Neg ScaleP
Scale
b. -ful
ScaleP
Scale
25/28
In tree structure
• ‘Lexical categories’ (like adjectives, verbs) fall into various
classes that can be described by various degrees of functional
structure
• Functional structure all the way down
• No “roots” in syntax
• Root makes sense as a morphological term — the base to which
affixes attach
26/28
To be continued...
27/28
References
Abney, Stephen. 1987. The english noun phrase in its sentential
aspect. MIT, Cambridge, Massachusetts: Doctoral Dissertation.
Aronoff, Mark. 1994. Morphology by itself. Cambridge: MIT Press.
Chomsky, Noam. 1957. Review of Skinner Verbal Behavior. Language
35. 26–58.
Ramchand, Gillian. 2008. Verb meaning and the lexicon. Cambridge,
Massachusetts: MIT Press.
28/28