Transcript Semantic and formal features in language change

Semantic and formal features
in language change
Elly van Gelderen
DGfS, March 2012
Outline
1 What are the Minimalist features?
2 How do they work in language change
and why?
3 Where do features `come from’?
4 Possible future directions.
The importance of features
Chomsky (1965: 87-88): lexicon contains
information for the phonological,
semantic, and syntactic component.
Sincerity +N, -Count, +Abstract...)
Chomsky (1995: 230ff; 236; 277ff):
semantic (e.g. abstract object),
phonological (e.g. the sounds),
and formal features:
intrinsic or optional.
The intrinsic ones are "listed explicitly in
the lexical entry or strictly determined by
properties so listed" (Chomsky 1995: 231)
and include categorial features, the Case
assigning features of the verb, and the
person and gender features of the noun.
Optional features are added arbitrarily and
are predictable from linguistic Principles
(e.g. nouns need Case or some kind of
licensing). They include the tense and
agreement features of verbs and the
number and Case features of nouns.
Features of airplane and build
(adapted from Chomsky 1995: 231)
airplane
build
semantic:
e.g. [artifact]
e.g. [action]
phonological: e.g. [begins with a vowel; e.g. [one syllable]
two syllables]
formal:
intrinsic
optional
intrinsic
optional
[nominal]
[number]
[verbal]
[phi]
[3 person]
[Case]
[assign accusative] [tense]
[non-human]
The "much more important distinction“
(1995: 277):
Formal features are: interpretable and
uninterpretable
airplane
build
Interpr. [nominal]
[verbal]
[3 person]
[assign
[non-human]
accusative]
Uninterpr [Case]
[phi]
Chomsky (2001: 10)
Phonological features are accessed at PF,
the semantic ones at LF, and the formal
ones accessible in the NS, but semantic
and formal “intersect”.
This intersection was not there in Chomsky
(1965: 142) where semantic features are
defined as not involved in the syntax.
The uninterpretable ones are valued and
only survive to PF; the interpretable ones
are relevant at LF.
Around 1998: AGREE
(1)
TP
T’
T
[u-phi]
DP
VP
many buffaloes V
[i-3] [i-P]
are
V’
PP
in the room
Formal features aren’t uniform
A)
Two-way, reciprocal: agreement [u-phi]
and [u-Case].
B) just interpretable: Cinque’s features for
modals and possibly [i-phi] in
Pronominal Argument Languages.
C) One-way, non-reciprocal: [u-neg] and
[i-neg]
So, `active’ is debatable as is the direction
(see e.g. Baker 2008). Carstens (2012):
“delayed valuation”: no (A).
Semantic and formal overlap:
Chomsky (1995: 230; 381) suggests: "formal
features have semantic correlates and
reflect semantic properties (accusative
Case and transitivity, for example)."
I interpret this: If a language has nouns with
semantic phi-features, the learner will be
able to hypothesize uninterpretable
features on another F (and will be able to
bundle them there).
CP and TP
(2)
VP
V
say
[u-ind]
CP
C’
C
that
[i-ind] she
[u-fin]
TP
T’
T
[i-past]
...
Why aren’t uninterpretable and
semantic enough?
Two reasons: (a) analytic languages may
not have uninterpretable features, and (b)
in (much) language change, we can see
all three at work.
Let’s therefore look at how these features
work in change.
Loss of semantic features
Full verbs such as Old English will with
[volition, expectation, future] features are
reanalyzed as having only the feature
[future] in Middle English.
And the negative
OE no/ne > ME (ne) not > -n’t
> ModE –n’t ... nothing, never, etc
Semantic > Interpretable >
Uninterpretable
(1) Ac nis nan scild trum[ra] wið ðæt ...
But NEG.is no shield stronger against the ...
`But there is no stronger shield against ...’
(2) ne ne helpeð nawiht eche lif to haben.
nor not helps not eternal life to have
`Nor does it help to have eternal life.’
(3) I can't do nothing for you either, Billy.
(4) No, I never see him these days
(BNC - A9H 350)
French Pronoun > Agreement
(well known)
(1)
(2)
Se je meïsme ne li di
Old French
If I myself not him tell
`If I don’t tell him myself.’ (Franzén 1939:20,
Cligès 993)
a.
Je lis et j'écris
I read and I-write
b.
*Je lis et écris
I read and write
c.
*Je probablement ai vu ça
I probably have seen that
(3)
(4)
(5)
(6)
J’ai vu
ça.
I-have
seen that
tu
vas où
2S go where
‘Where are you going?'
Moi, je ....
me, I ...
si un: un Russe i va en France
Swiss
if a Russian il goes to France
‘If a Russian goes to France.’
(Fonseca-Greber 2000: 335)
Old French
Emphatic Regular
Subject
je/tu
Oblique
moi/toi
>
Modern French
Emphatic Regular
zero
moi/toi
je/tu
me/te
moi/toi
me/te
The cycle of phi-features
noun > emphatic > pronoun > agreement >
[sem]
[i-phi] [i-phi]/[u-phi] [u-phi]
0
Demonstrative to copula
Assume copulas have:
be
remain
seem
[i-loc]
[i-loc]
[i-loc]
[i-ASP]
[i-M]
Source for [loc]? Verbs and demonstratives
D
[i-loc]
[i-phi]
[u-T]
> copula
> [i-loc]
> [u-phi]
> zero
> --
Chinese
Old Chinese shi ‘this’
(1) fu
yu gui
shi ren
Riches and honor
this men
zhi suo
yu
ye
GENNOM
desire
BE
‘Riches and honor, this is men’s desire.’
(2) Shi shi lie
gui
this is
violent
ghost
‘This is a violent ghost.’
Modern Chinese
(3)
(4)
Zhe shi lie
gui
this BE violent
ghost
‘This is a violent ghost.’
Shi wo de
cuo
be 1S POSS
fault
‘It’s is me (who is) at fault.’
(1)
(2)
(3)
Old Egyptian (1) > Middle (2)
a. rmt p-n
man MS-PROX
`this man.’
b. ntr-w
jp-w
god-P
MP-DIST `those gods.’
̩tmj-t
pw jmn-t
city-F
be west-F
`The West is a city.’
(Loprieno 1995; 2001)
p
-w
>
pw
[i-3MS] [i-distal]
[i-loc]
Other explanations
The elephant that happy
TOPIC
SU VP
↓
SU
copula
VP
The question would be why first (or second)
person pronouns are never reanalyzed as
copulas since they are frequent topics.
Demonstrative to article and
complementizer
(1)
ðis gære for þe king Stephne ...
this year went the King ...
(2) monig oft gecwæð þæt te suð ne
norð ... oþer ... selra nære
many often said that that south nor north,
other better not-was
`It was often said that no better one could
be found North or South.' (Beowulf 858)
Features of the English DP
(1)
(2)
a.
b.
c.
*That the dog loves their the toys.
I saw that.
*I saw the.
DP
DP
that
D’
D
NP
[i-loc] D
NP the
3S
[i-ps]
3S [u-phi]
Demonstrative
[i-3S]
[i-loc]
article
[u-phi]
(1)
complementizer
[u-T]
(copula)
[i-loc]
Mi da i
tatá
Saramaccan
I
be your father
‘I am your father.’ (McWhorter 1997)
C [iT] > [uT]
Pesetsky & Torrego (2001): that is the spell-out of
a T with interpretable tense features; the finite C
has tense features that must be checked by
either a nominative, by that, or by an auxiliary.
Two phenomena are explained: the optionality of
that in English complement clauses (since either
the subject DP or that can check [u-T] of C) and
the that-trace effect in Modern English.
Problems: the lack of C-deletion in Old English and
no `that-trace' effects. Van Gelderen (2011)
argues that the lack of C-deletion is due to the
interpretable features of C in the older period. As
they are reanalyzed as uninterpretable, C
becomes deletable.
Latin: From Neg to Q
(1)
tu-ne
id veritus
es
you-Q
that fear
be
`Did you fear that?’ (Greenough et al.
1931: 205)
Negatives value the [u-Q] of the PolP
through their [i-neg]; “if the negative quality
somehow weakens, it is reanalyzed as a
PolP head whose polarity is not specified.”
(van Gelderen 2011: 295).
Where do features come from?
The Minimalist program has shifted the
emphasis from UG to third factors and
from syntactic parameters to lexical ones,
i.e. features. One of the reasons to
deemphasize UG is the supposed lack of
evolutionary depth.
Third factors, however, are vague and
feature theory is not well-developed.
Three Factors
“(1) genetic endowment, which sets limits on the
attainable languages, thereby making language
acquisition possible; (2) external data, converted
to the experience that selects one or another
language within a narrow range; (3) principles
not specific to FL [the Faculty of Language].
Some of the third factor principles have the
flavor of the constraints that enter into all facets
of growth and evolution.... Among these are
principles of efficient computation”. (Chomsky
2007: 3)
From early GenGr to Minimalism
Universal Grammar
UG and Third factors
+
>>
+
Input
Input
(Scottish English, Western Navajo, etc)
=
=
I-language
I-language
E-language
Borer-Chomsky-Conjecture
"All parameters of variation are attributable
to differences in the features of particular
items (e.g., the functional heads) in the
lexicon." (Baker 2008: 156)
Muysken (2008: 6):
“I find the generative literature on functional
categories rather vague.”
Cinque and Rizzi (2008) discuss the
question of the number of functional
categories. There are 32 in Cinque (1999:
130) and around 40 in Kayne (2005).
Cinque and Rizzi, using Heine & Kuteva’s
2002 work on grammaticalization, come
up with 400 features that are targets in
Heine & Kuteva.
Benincà & Munaro (2010: 6-7) note that
syntax has reached the detail of
phonological features.
Cf. phonology:
“The main task of feature theory, then, is to
find the phonetic features which accurately
describe the attested phonologically active
classes in the world’s languages”
(Samuels 2012: 4).
The first 20 features in
Heine and Kuteva
permissive, possibility, agent, comparative,
material, partitive, past/near, A-possessive,
since (temporal), superlative,
complementizer, dative, infinitive, patient,
purpose, temporal, until (temporal), only,
NP-and, subordinator
Some other questions
Muysken (2008: 46): “features which are doubly
expressed ... but receive a single interpretation,
must be functional.”
Which feature can value which?
[u-phi] is easy, as long as it gets a value such as
from i-3, i-P etc.
[u-pol]:
i-neg?
[u-T]:
i-past?
[u-ind]:
i-ind?
How about the order of categories?
Chomsky (2001: 12):
“Assume that substantive categories are
selected by functional categories. V by a
light verb, T by C”.
Cinque Hierarchy?!
Challenge: acquisition of features
and their order
Jackendoff (2002), based on Bickerton
(1990), suggests that pre-linguistic primate
conceptual structure may already use
symbols for basic semantic relations. This
may include spatial and causal concepts.
“Agent First, Focus Last ... are `fossil
principles’ from protolanguage”. Homo
erectus (1 million BP) may have had
protolanguage. This gives the innate
faculty longer to incorporate this.
The acquisition of semantic
features
Chomsky (1965: 142): “semantic features ...
too, are presumably drawn from a
universal ‘alphabet’ but little is known
about this today and nothing has been
said about it here.”
Chomsky (1993: 24) vocabulary acquisition
shows poverty of the stimulus.
The status of meaning, i.e. sem features
“Les idées ... ne tirent en aucune sorte leur
origine des sens ... Notre ame a la faculté
de les former de soi-même.”
`Ideas do not in any fashion have their origin
in the senses ... Our mind has the faculty
to form those on its own.’ (Arnauld &
Nicole 1662 [1965]: 45)
How to address the PoS
• Pinker (1984: 57): categorization < semantic
properties and Lebeaux (1988: 44): grammatical
categories are centered in cognitive ones.
Where do semantic and cognitive categories
come from? UG?
• Geach (1957: 22-23): “Abstractionists rarely
attempt an abstractionist account of logical
concepts, like those of some, or, and not” ... “In
the sensible world you will find no specimens of
alternativeness and negativeness from which
you could form by abstraction the concept of or
or of not”.
Acquisition: sem > [i-F]/[u-F]
(1)
(2)
(3)
(4)
(5)
(6)
like a cookie (Abe, 3.7.5)
no the monster crashed the planes down like
this like that (Abe, 3.7.5)
I wan(t) (t)a show you something # I mean
like this thin ? (Abe, 3.7.5)
I feel like having a pet do you? (Abe, 4.8.20)
watch it walks like a person walks.
(Abe, 4.9.19)
Daddy # do you teach like you do [//] like
how they do in your school? (Abe, 4.10.1)
Do we need uninterpretable?
-Two negative cycles:
A) Using an indefinite, e.g. nothing/never/a
bit in English, French, Arabic
B) Using a new verb, e.g. Chinese
-Languages without overt agreement
Are there features for External
Merge?
• V [u-Theme] (or u-c/u-m)?
• Although the V seems the probe, it cannot
have uninterpretable features! Can we do
away with theta-roles? Not for theme!
Conclusions
Recent shift towards third factors and
parametric features: we need to be careful
how many mechanisms we allow.
All change is in the lexicon: sem>i-F>u-F
What does the Poverty of the Stimulus
argument mean for vocabulary
acquisition?
References
Adger, David & Peter Svenonius 2010. Features in
Minimalist Syntax. ms
Benincà, Paola & Nicola Munaro 2010. Introduction. In
Benincà, Paola & Nicola Munaro (eds), Mapping the Left
Periphery, 3-15. OUP.
Chomsky, Noam 1993. Language and Thought.
Chomsky, Noam 1995. The Minimalist Program. MIT Press.
Chomsky, Noam 2001. Derivation by Phase.
Chomsky, Noam 2007. Approaching UG from below, in Uli
Sauerland et al. (eds), Interfaces + Recursion =
Language, 1-29. Berlin: Mouton de Gruyter.
Cinque, Guglielmo 1999. Adverbs and Functional Heads.
Oxford: Oxford University Press.
Cinque, Guglielmo & Luigi Rizzi 2008. The cartography of
syntactic structures V. Moscati, ed. CISCL Working
Papers on Language and Cognition, 2, 43-59.
• Geach, Peter. 1957 Mental Acts.
• Gelderen, Elly van 2011. The Linguistic Cycle. OUP
• Heine, Bernd & Tania Kuteva 2002. World Lexicon of
Grammaticalization. Cambridge: Cambridge University
Press.
• Jackendoff, Ray 2002. Foundations of language. Oxford.
• Lebeaux, David 1988. Language acquisition and the
form of the grammar. Doctoral dissertation, University of
Massachusetts, Amherst.
• Muysken, Pieter 2008. Functional Categories. CUP
• Panagiodis, E. Phoevos 2008. Diachronic stability and
feature interpretability. In Theresa Biberauer (ed.) The
Limits of Syntactic Variation. Amsterdam: John
Benjamins.
• Pesetsky, David & Esther Torrego 2007. The Syntax of
Valuation and the Interpretability of Features. In Simin
Karimi et al. Phrasal and Clausal Architecture, 262-294.
Amsterdam: John Benjamins..
• Pinker, Steven 1984. Language Learnability and
Language Development. Cambridge: Harvard University
Press.
• Samuels, Bridget 2012. The Emergence of Phonological
Forms. ms
• Shlonsky, Ur 2010. The Cartographic Enterprise in
Syntax. Language and Linguistics Compass 4/6: 417429.
Even more complex
Pesetsky & Torrego (2006) argue that
valuation and interpretability are
independent. Unvalued features act as
probes and these can be either
interpretable or uninterpretable. The tense
features in T are interpretable but
unvalued whereas the tense features on
the verb are uninterpretable but valued.
Unidirectional or not?
Richness:
For instance, Panagiodis (2008: 447) states
that diachronic processes may introduce
or eliminate uninterpretable features and
rearrange them into new combinations.
Change suggests:
Most changes are unidirectional.