Lingue senza aggettivi? - Grandionline.net

Lingue senza aggettivi?
Mark Post (2008), Adjectives in Thai: Implications
for a functionalist typology od word classes,
Linguistic Typology, 12:3, 339-381
Lingue del Sud Est asiatico: senza aggettivi?
No, se si assume l’idea che le parti del discorso
siano universalmente definite essenzialmente in
base alla condivisione di funzioni pragmatiche

Adjective classes are nowadays routinely identified as prototypically
containing words denoting property concepts (p. 341)
concetti graduabili: aggettivi
concetti non graduabili: nomi e verbi
Fundamentally, the semantic property of gradability would seem to lend
adjectives an aptness for degree modification (more / less, a lot / little,
the most / least), for lexicalization as points on a property scale (tiny,
small, big, huge,…), as well as for comparison in terms of the relative
degree of properties exhibited by two referents (p. 341)

Dixon, R. M. W. (1977), Where have all the adjectives gone?, “Studies in
Language”, 1, 19–80
Dixon, R. M. W. / Aikhenvald, A. Y. (eds.) (2004), Adjective Classes. A Cross-
Linguistic Typology, Oxford, Oxford University Press
CORE:
Dimension (big, small…)
Age (new, old…)
Value (good, bad…)
Colour (black, white…)
PERIPHERY:
Physical property (hard, soft…)
Human propensity (happy, sad…)
Speed (fast, slow…)
Functionally, adjectives prototipically attribute features to a referent

Adjectives are often viewed as inherently mono-relational, in the semantic sense
that their denotation may be viewed as incomplete without simultaneous construal
of some entity or set which exhibits the stated property (e.g. a man who is tall), and
in the syntactic sense that they either project (as predicate), cooccur with (as
copula complement), or depend on (a modifier) one and only one referring
expression.
Thai
Terms denoting properties are primarily divisible into
-Verblike forms (the vast majority)
- Nounlike forms

Verblike properties include terms from all of the core and peripheral semantis fields:
dii ‘good’
jεεÎ ‘bad’
rO´On ‘hot’
naÍaw ‘cold’
Nounlike property concepts mainly correspond to the fields of PHYSICAL PROPERTY,
HUMAN PROPENSITY, and COLOUR:
thammachaÎat ‘natural(ness)’
sùantua ‘private/privacy’
saniÍm ‘rust(y)’

Proprietà distribuzionali
Verblike property concepts occur as intransitive predicate and do not occur as
copula complement
Khon níi dəən
CLF:person PRX walk
‘This person walks’ (active verb, intransitive predicate)
*Khon níi pen dəən
CLF:person PRX ACOP walk
‘This person walks’ (active verb, intransitive predicate)
Khon níi dii
CLF:person PRX good
‘This person is good’ (verblike property term, intransitive predicate)
*Khon níi pen dii
CLF:person PRX ACOP good
‘This person is good’ (verblike property term, intransitive predicate)
PRX = proximate demonstrative
ACOP = attributive copula

Nounlike property concepts behave more like concrete and abstract nouns in
occurring as complement of an attributive copula
Khan níi pen sanǐm
CLF:VEHICLE PRX ACOP rust(y)
‘This vehicle is rusty’ (nounlike property term, copula complement)
*Khan níi sanǐm
CLF:VEHICLE PRX rust(y)
Khon níi pen phará?
CLF:PERSON PRX ACOP monk
‘This man is a monk’ (concrete noun, copula complement)
*Khon níi phará?
CLF:PERSON PRX monk

Verblike property terms take direct verbal predicate negation in mâj ‘NEG’ whereas
nounlike property terms, like nouns, again usually require support of a copula or
other predicative functor.
Ma c’è una costruzione in cui possano occorrere solo i cosiddetti verblike property
terms o i cosiddetti nounlike property terms? Solo in questo caso potremmo
considerarli una classe autonoma…
All and only verblike property terms may uncontroversially occupy the position
provisionally marked “x” in the bare comparative of a discrepancy construction;
stative and active verbs cannot usually occur in this position
[NP1] [X] [COMP] [NP2]

Khon níi sǔuN kwàa khon nân
[CLF:PERSON PRX] NP1 [tall] X [more] COMP [CLF:PERSON DST] NP2
‘This person is taller than that person’ (verblike property term)
*khon níi khít/dəən kwàa khon
[CLF:PERSON PRX] NP1 [think/walk] X [more] COMP [CLF:PERSON
nân
DST] NP2
‘This person thinks/walks more than that person (does)’ (stative/active verbs)
DST = Distal demonstrative
Such distributional differences [can be used] as evidence for
the coalescence of a distinct class of lexemes whose semantic
core is made up of property terms.

Khon níi sǔuN thâw khon nân
[CLF:PERSON PRX] NP1 [tall] X equal [CLF:PERSON DST] NP2
‘This person is tall as that person (is)’ (verblike property term)
*khon níi khít thâw khon nân
[CLF:PERSON PRX] NP1 [think] X equal [CLF:PERSON DST] NP2
‘This person thinks/ as much as that person’ (stative/active verbs)
DST = Distal demonstrative

Una parte del discorso non può essere chiusa: se esiste una classe di aggettivi,
deve esistere anche il modo per formare nuovi aggettivi.
nâa- “of an entity, have/be of a quality such that it seems like it would be nice to
experience interaction with in terms of VERB”
thanǒn sên nân nâa-dəən kwàa sên níi
road CLF:LINE DST AZR-walk more CLF:LINE PRX
‘That road looks more walkable (better or more suitable to walk on) than this one’
châan- “having the characteristic of doing”
khîi- “having the negative characteristic of doing”

“There is a class of terms in Thai which closely resembles the adjective classes of
many other languages in terms of semantic contents, internal structure, and
distribution relative to other lexical classes, and this class should therefore bear
the label “adjective””
Discovery of a linguistic universal “adjectives, properly defined, requires
establishment of no single necessary and sufficient formal criterion such as
“occurs as copula complement”; what it requires is the discovery that no language
fails to develop grammatical means of treating property concept words differently
from other types of terms. The typological-oriented prediction of Dixon (2004) that
adjectives (so defined) will be found in every language can be falsified through
discovery of a language in which property concept words are in fact NOT treated
in any way differently from another well-defined type of term (say, verb or noun).

Strutture sintattiche e cervello: esistono lingue impossibili?

Area di Broca

“L’ipotesi più probabile è che i limiti delle sintassi delle lingue umane siano dovuti
a una matrice di stampo biologico”
(Andrea Moro (2006), I confini di Babele. Il cervello e il mistero delle lingue
Impossibili, Milano, Bompiani, p. 146)
Perché non ci sono mai (o quasi) errori sintattici?
Perché un bambino può dire, anzi dice io ando o io piangio ma non dice mai casa la?
Esiste una guida biologicamente determinata per l’apprendimento di una lingua,
determinata dall’architettura funzionale del cervello?

1° esperimento: la autonomia della sintassi e il suo ‘posto’ nel cervello
Una lingua immaginaria:
- Hanno disbato le artine
- Molte grapotte amionarono
- Molti celuci furono taffivati
- Nessun cribaso è stato incenghito
…
‘Errori’ fonologici:
- Hanno dinsbato le artine
- Molte grapotrte amionarono
- Mosnti celuci furono taffivati
- Nessun cribaso è stgtato incenghito
…

‘Errori morfosintattici’
- Hanno disbata le artine
- Molti grapotti sono stata amionati
- Molti celuci fu taffivati
- Nessun cribaso siamo incenghito
…
‘Errori’ sintattici
- Hanno disbate artine le con gli ziggoli
- Grapotte molte amionarono
- Celuce delle furono taffivate
- Cribaso è incenghito nessuno a rimbaudo
…
“La nostra speranza […] era di vedere attivare zone diverse a seconda del tipo di
errore e, in defintiva, di vedere per la sintassi una zona diversa dalle altre”
(Andrea Moro (2006), I confini di Babele. Il cervello e il mistero delle lingue
Impossibili, Milano, Bompiani, p. 185)

“Solamente nel caso dell’errore di tipo sintattico si attivano alcune zone sottocorticali
dell’encefalo insieme alla componente profonda dell’area di Broca […].
Questo è il punto centrale. Il riconoscimento dell’errore di tipo sintattico coinvolge una
rete complessa che non si riscontra negli altri tipi di errori e tale rete non è
rappresentata in un’unica area corticale ma si presenta come un insieme integrato di
zone diverse”
(Andrea Moro (2006), I confini di Babele. Il cervello e il mistero delle lingue
Impossibili, Milano, Bompiani, p. 189)
“Non solo dunque non esiste un’area singola per il linguaggio […], ma non esiste
nemmeno un’area singola per la sintassi!”
(Andrea Moro (2006), I confini di Babele. Il cervello e il mistero delle lingue
Impossibili, Milano, Bompiani, p. 190)

2° esperimento: “far apprendere a dei soggetti adulti delle lingue straniere,
«nascondendo», tra le regole delle grammatiche che i soggetti si apprestavano
a imparare, delle regole che violano la grammatica universale, più specificamente
delle regole che violano la dipendenza dalla struttura”
(Andrea Moro (2006), I confini di Babele. Il cervello e il mistero delle lingue
Impossibili, Milano, Bompiani, p. 195)
(Andrea Moro (2006), I confini di Babele.
Il cervello e il mistero delle lingue
Impossibili, Milano, Bompiani, p. 203)

- Dipendenza dalla struttura
Es. La ragazza di Pietro suona bene il pianoforte
*Il Pietro pianoforte bene di ragazza suona la
*Il pianoforte di Pietro suona bene la ragazza
*il gatto si nascondono dietro l’albero
I bambini che hanno inseguito il gatto si nascondono
dietro l’albero

“Il fatto di rispettare o meno la dipendenza strutturale è dunque irrilevante per quanto
riguarda l’accuratezza dell’apprendimento: sia che si trattasse si regole possibili, sia
che si trattasse di regole impossibili, i soggetti arrivavano a una padronanza del tutto
comparabile”
(Andrea Moro (2006), I confini di Babele. Il cervello e il mistero delle lingue
Impossibili, Milano, Bompiani, p. 205)
Non ci sono lingue che non si possono imparare!
Ma ci sono lingue che, naturalmente, non si sviluppano mai,
anche se, appunto, saremmo in grado di apprenderle o usarle…

“Al crescere dell’accuratezza delle risposte sui giudizi di grammaticalità […]
l’attività dell’area di Broca […] aumenta per le regole possibili mentre diminuisce
per quelle impossibili. Il cervello ha, per così dire, «smistato» i dati sintattici (senza
che i soggettine avessero coscienza) e ha fatto elaborare solo le frasi che preservano
la dipendenza dalla struttura dall’area normalmente predisposta per i compiti sintattici
(l’area di Broca); nell’elaborazione di frasi che non rispettano la dipendenza dalla
struttura, invece, l’attività nella stessa area diminuisce progressivamente”
(Andrea Moro (2006), I confini di Babele. Il cervello e il mistero delle lingue
Impossibili, Milano, Bompiani, p. 205)

Perceiving and naming actions and objects
M. Liljeström, A. Tarkiainen, T. Parviainen, J. Kujala, J. Numminen, J. Hiltunen,
M. Laine, and R. Salmelina
NeuroImage 41 (2008) 1132–1141
The existence of patients in whom focal brain damage is combined with
selective impairments of grammatical categories has led to the suggestion that
there are distinct cortical areas for processing verbs and nouns
Lesion data has implied a link between left frontal areas and verb
processing, whereas noun processing seems to depend on the integrity of
left temporal areas

Nouns and verbs are distinguished in the brain on the basis of their semantic
properties, nouns typically referring to objects and verbs to actions.
In the present fMRI experiment, healthy subjects were asked to silently name
actions and objects that were presented as simple line drawings. We used two
sets of pictures, one depicting actions performed on/with objects and the other
displaying only the objects. Subjects named both actions and objects from the
action images, and objects from the object images. Actions were not named
from pictures with isolated objects, as this task would require the subject to
make inferences about the actions, a process not required in object naming. This
design addresses the following questions:
(i) Do action and object naming activate different cortical regions when the
stimulus is identical?
(ii) How does the content of the image (with/without action) modulate the brain
correlates of object naming?

The task was to silently name actions or objects from simple line drawings.
Two sets of images were used, each with 100 scenes. Action images illustrated
a simple event (e.g. to write with a pen) whereas object-only images consisted
of objects from the same images when the action had been dissolved into
arbitrary lines in the background, in order to keep the visual complexity of
the image unchanged.
The corresponding object and action words were mostly of medium to high
frequency in the Finnish language.
The verb and noun corresponding to one image always had different word
stems.

The experiment consisted of three conditions:
(i) action naming from action images (Act),
(ii) object naming from action images (ObjAct),
(iii) object naming from object-only images (Obj).
Task periods (30 s) and rest periods (21 s) alternated in a block design. There
were two sessions, each lasting about 13 min.
In order to avoid movement artifacts the participants were instructed to name the
actions or objects silently. Subjects were asked to keep their eyes straight ahead
during the rest condition and not to move during the experiment.
Results: A similar network of cortical areas was activated in all three conditions,
including bilateral occipitotemporal and parietal regions, and left frontal cortex.

The occipitotemporal cortex and the fusiform gyrus were activated bilaterally in all
conditions. When actions or objects were named from action images (Act and
ObjAct) the activation additionally encompassed the left posterior middle
temporal cortex. Activation was seen in the left inferior frontal gyrus in all tasks. In
both action and object naming from action images (Act and ObjAct) the activation
centered in the operculum of the inferior frontal gyrus and extended superiorly to
include the precentral gyrus and anteriorly to BA 47. The frontal activation was
less pronounced when objects were named from object-only images (Obj), mainly
encompassing the inferior frontal region.

Naming objects from the images with action context compared to naming the same
objects from object-only images (ObjActNObj, Fig. 3B) resulted in significantly
stronger activation in a large mostly left-lateralized network, including the precentral
gyrus (BA 6/9), the inferior frontal gyrus (pars opercularis and triangularis), the
inferior and superior parietal lobules (BA 7/40), and the posterior middle temporal
gyrus. Naming actions compared to naming objects from object-only images
(ActNObj, Fig. 3C) revealed enhanced activation in the left supramarginal gyrus
(parieto-temporal junction; BA 40), left and right posterior middle temporal cortex,
left precentral gyrus and superior medial frontal gyrus.
Our results indicate that action and object naming engage a common cortical
network, but that the provided input (action pictures, object-only pictures) and the
requested output (verb, noun) influence the level of activation in a subset of areas
within that network.
The stronger activation in naming objects from pictures with action context,
relative both to naming actions from the same images and to naming objects
from object-only pictures, suggests involvement of additional task-specific
processes.

We found no regions specific to nouns as a grammatical category. Our results
agree with a recent study showing that when naming events either as verbs or
nouns from identical images, no differences were found between verbs or nouns
as grammatical categories.
Our results converge with previous evidence showing that retrieval of verbs and
nouns in the healthy human brain using identical stimuli in a picture naming task
engages a similar distributed cortical network, as measured with BOLD fMRI.
Importantly, however, the content of the image (action vs. object only) had a
pronounced effect on the activation in parts of that network that have previously
been implicated in processing of action knowledge. Furthermore, object naming in
the context of action revealed additional activations, both in comparison to verb
retrieval from the same set of images, and in comparison to noun retrieval from
images not depicting action, suggesting that attention may be more directed
towards motor-based properties of objects when they are presented not as single
entities but as part of images that also depict the relevant action.