PALOMAR: A Computer-aided Analysis of some Lexical and Stylistic Features

Abstract
This paper presents a few initial results of a study
which is being carried out at the Institute of Computational Linguistics in Pisa, fitting into the framework of computer-based approaches to literary
text analysis. The object of this investigation is
Palomar, a work of the well-known Italian writer
Italo Calvino. This book, published in 1983, is one
of Calvino’s later works and is very representative
of the author’s stylistic and thematic aspects. The
paper focuses on the lexical structure and the
particular use of the different categories on the
syntagmatic level (i.e. the way they combine), as
well as the computer tools used to exploit this
information in a consistent and exhaustive way.
1. The Pisa tools for text analysis
In the last few years, at ILC in Pisa, a set of tools,
known as the PiSystem, has been developed and
continuously improved in order to make possible
the exploitation of text corpora according to the
most varied user needs. In particular, a statistically
based procedure has been implemented for the
automatic lemmatization and tagging of Italian
texts: the Pitagger (Picchi 1994). This procedure
assigns the grammatical category and the lemma
to the word forms occurring in whatever Italian
text is being analyzed. Obviously, the output of
this fully automatic procedure is not completely
correct. Furthermore, if a deep morphological analysis is requested, a choice must be made among,
for example, the homograph inflected verb forms.
So, in order to obtain a correct and sufficiently
complete text analysis, as in the case of studies
aiming to individuate and describe stylistic features, an interactive procedure, called Tagghand,
helps the user to correct and to deepen his/her
research. After that, another interactive procedure,
the Dblemcor, allows the user to make new changes on the already lemmatized word forms. This
is a very useful tool because this kind of work is
subject to doubts and afterthoughts. The difficulty
of choice between different grammatical categories when dealing with borderline cases such as
past participle / adjective, infinitive / noun, present
participle / adjective, present participle / noun and
so on is well known by anyone who has ever done
a lemmatization.
In the work presented here both procedures (the
Tagghand and the Dblemcor) were used and a
fine-grained coding was performed to allow further and more sophisticated researches relating to
grammatical categories, their employ and their
sequential combinations, from a stylistic point of
view. To this end, procedures offering the possibility of selecting the lemmata according to either
grammatical categories or morphological codes,
besides the various ending strings, turned out to be
very helpful. In the annexed figures a few examples of different subsets of this lexicon are shown.
They were obtained by querying for:
– the feminine singular nouns derived by the
suffix -ezza, Fig. 1;
– the masculine singular adjectives derived by
the suffix -ico, Fig. 2;
– the inflected verbal forms carrying as morphologic code “S3IP”, Fig.3;
– the word forms carrying as grammatical code
“W”, i.e. the proper nouns, Fig. 4.
2 Lexical features
According to unanimous critical judgement, Calvino’s writing is characterized by great clarity,
accuracy and punctuality, which should be reflected in his lexical choices (see Mengaldo 1991 and
1994). So, in the survey on Palomar’s lexical
structure, the use of the computer appeared particularly useful to verify and quantify, when possible, critical intuitions and suggestions. In the limited space of this presentation only a few
features will be considered, in particular:
– the creation of new words to outline particular ideas, minds or sensations;
– the preference accorded to certain types of
word formations, and, vice-versa, the limited
use of others (e.g. the diminutives).
As far as the first feature is concerned, the data
extracted from the text confirms many critical
remarks asserting that, normally, the creativity of
this author is not expressed by means of new
words, but rather relies on an exceptional linguistic competence (also as far as scientific sublanguages are concerned) joined with a particular ability
to stress all the possibilities of those syntactic
combinations which the Italian language plentifully allows for. This seems to be true also for Palomar. In fact, the new words, which constitute a
little subset within this lexicon, are mostly original
matchings of common words by hyphenation, and,
if we do not consider those adjectives expressing
different color shades, such as bronzeo-rosea
(bronze-rosy) or nero-giallastro (black-yellowish), we have fourteen new compounds in all,
aiming to express particular thoughts, opinions
and feelings of the writer, listed and literally translated below:
braccia-zampe (arms-paws)
continenti-cimiteri (continents-cemeteries)
divergenza-convergenza (divergence-convergence)
finestra-vetrina (window-shop window)
gabbia-vetrina (cage-shop window)
lumpen-pennuti (lumpen-feathered)
spiaggia-immondezzaio (plage-garbage)
silenzio-discorso (silence-speech)
umanità-sabbia (humanity-sand)
uomo-albero (man-tree)
uomo-bue (man-ox)
iperlavoro (hyperwork)
ultralavoro (ultrawork)
Regarding the second feature, the lists of nouns
and adjectives (extracted according to their inverse alphabetical order) prove that diminutives, augmentatives and other types of alterations, belonging to an emotional more than a precise and
punctual language, are scarcely used. In the nouns
list only the following alterated word-forms appear:
farfallina (little butterfly)
macchiolina (little spot)
fogliolina (little leaf)
pianticelle (little plants)
cascatelle (little waterfalls)
lombatine (little sirloins)
scimmione (big monkey)
fiorellini (little/nice flowers)
pilastrini (little pilasters)
ragazzotti (big/rough boys)
praticello (little/nice lawn)
bastoncino (little stick)
The adjectives subset shows a few superlatives
and four alterates in -ino and -astro:
velocissima (very quick)
elaboratissima (very elaborate)
rapidissime (very rapid)
bellissime (very beautiful)
vicinissimi (very near)
nitidissimo (very clear)
bianchissimo (very white)
difficilissimo (very difficult)
minutissimo (very minute)
biancastra (whitish)
giallastro (yellowish)
freddina (chilly)
giallini (faded yellow)
The possibility of selecting subsets of lemmatized
word forms in inverse alphabetical order, according to either their syntactic categories or to a
particular derivative suffix, evidenced a great variety and richness in the use of derived nouns and
adjectives which appear to be, on the lexical level,
Calvino’s main way of obtaining accuracy and
exactitude.
Because it is impossible here to draw up complete
lists of them, we will provide some quantitative
data and we will comment on only a few of the
more interesting and widely used examples. The
analysis of the complete list of the inverse-ordered
nouns shows that the most frequent derivative
suffixes attested within this lexicon are the following:
-sione(i)/-zione(i) 205
-ità/età/tà 80
-anza(e)/-enza(e) 79
-mento(i) 60
-ezza(e) 36
-sura(e)/-tura(e) 31
A first division of these suffixes can be made and
we can state that three of them (-zione, -mento,
-tura) are deverbal and derive actions or processes
nouns. The other three suffixes (-ità, -anza, -ezza)
are deadjectival and give rise to nouns expressing
qualities and conditions. In looking at these sets of
derived nouns, we can confirm a typical character
attributed to Calvino’s writing: many words belong to scientific sublanguages, testifying to his
strong interest in the natural, physical, mathematical and anthropological sciences. These terms
abound, especially within the first group of deverbal nouns while, in the second group, scientific
terms are less frequent, even if a few terms typical
of physical and chemical fields were found, as can
be seen from the lists below:
corrosione (corrosion)
dispersione (dispersion)
dissociazione (dissociation)
disgregazione (disintegration)
assimilazione (assimilation)
inclinazione (inclination)
migrazione (migration)
pigmentazione (pigmentation)
distribuzione (distribution)
classificazione (classification)
variazione (variation)
angolazione (angle)
proliferazione (proliferation)
segmentazione (segmentation)
intersezione (intersection)
induzione (induction)
condizionamento (conditioning)
comportamento (behaviour)
procedimento (process)
mutamento (change)
ragionamento (reasoning)
esperimento (experiment)
punteggiatura (punctuation)
nomenclatura (nomenclature)
temperatura (temperature)
partitura (score)
scrittura (writing)
struttura (structure)
solidità (solidy)
fluidità (fluidity)
potenzialità (potentiality)
granulosità (granularity)
continuità (continuity)
umidità (humidity)
profondità (depth)
233
densità (density)
luminosità (luminosity)
discontinuità (discontinuity)
discrepanza (discrepancy)
tolleranza (tolerance)
sequenza (sequence)
trasparenza (transparency)
esperienza (experience)
consistenza (consistency)
lunghezza (length)
altezza (height)
completezza (completeness)
ampiezza (amplitude)
lucentezza (brilliancy)
compattezza (compactness)
Another interesting point evidenced by the simple
survey on the lexicon composition is that some
scarcely productive types of suffixes can also be
relevant in the research of the links between lexicon and style because they are carriers of typical
features. This is the case with the deadjectival
suffix -ore(i) and the denominal suffix -ìo(ìi). The
first is employed in the semantic field of colours
and visual perception, the second carries a particular aspectual and intensive value in both the
fields of vision and audition. So they are clues of
a central theme in this work: the mechanisms of
perception (in particular the sight) as source of
knowledge; Palomar being an attentive, careful
and nearly maniacal observer of the world surrounding him. Besides, they confirm a critical
suggestion by Mengaldo about a certain preference for the derivatives in -ìo. Below are a few
examples literally (when possible) translated:
biancore (whitheness) brontolìo (grumbling)
bagliore (flare) cinguettìo (twittering)
splendore (splendour) borbottìo (mumbling)
fulgore (splendour) chioccolìo (gurgling)
grigiore (greyness) tremolìo (twinkling)
chiarore (glimmer) scintillìo (glitter)
Obviously, the list of the word frequencies is the
first way to individuate thematic features and, in
the Palomar frequencies list, we find that sguardo
(look) is, after mondo (world) and cielo (sky), the
third more frequent semantic word. Nevertheless,
it is interesting to detect similar indications also at
different linguistic levels, such as the derivatives
word classes.
Remarks and comments of the same kind can be
done for the adjectives. They also present many
derivatives typical of scientific languages (-bile,
-ico) and some others which are more literary or
latinizing (see -eo for example). The deverbals
(-bile, -ente, -ivo), and the denominals (-ico, -oso,
-eo) were among the most frequent suffixes found.
Below quantitative indications and a few examples are given:
-ico(a,he,i) 100
-bile(i) 66
-ente(i) 57
-oso(a,e,i) 52
-ivo(a,e,i) 34
-eo(a,e,i) 31
probabile (probable) variabile (variable)
tangibile (tangible) visibile (visible)
sensibile (sensible) osservabile (observable)
cosmico (cosmic) biologico (biological)
allegorico (allegorical) simmetrico (symmetric)
geometrico (geometric) simbolico (symbolic)
erboso (grassy) granuloso (granulous)
cremoso (creamy) poroso (porous)
gustativo (gustative) olfattivo (olfactory)
auditivo (auditive) visivo (visual)
erbaceo (herbaceous) bronzeo (bronze)
aereo (airy) argenteo (silvery)
vitreo (vitreous) corporeo (bodily / corporeal)
roseo (rosy) omogeneo (homogeneous)
latteo (milky) sidereo (sidereal)
3. The way Calvino combines the words:
the adjectivation
As stated above, the other principal way used by
Calvino to reach extraordinary accuracy and flexibility in his writing is through the use of a ductile
and varied combination of grammatical categories
within the sentences.
The Italian language allows, in this sense, great
possibility of variation, and the adjectivation in
Palomar is a perfect exemple of this. A procedure
which allows the retrieval of grammatical combinations evidenced many schemes of noun-adjective association. Below the more frequent patterns
are listed, together with some relative examples
literally translated:
- type ASA (ADJ+N+ADJ)
minuta devastazione erbacea(minute devastation herbaceous)
lunga notte biologica (long night biological)
grande gorilla nera (big gorilla black)
perpetua agitazione ansiosa (perpetual agitation anxious)
esatto equivalente grafico (exact equivalent graphic)
aguzzo triangolo dorato (pointed triangle gilt)
opaco colore serale (opaque colour evening)
solito grigiore quotidiano (usual greyness daily)
- type SAeA (N+ADJ+CONJ+ADJ)
realtà informe e dissennata (reality shapeless and mad)
attenzione minuziosa e (attention meticulous and
prolungata prolonged)
precisione nomenclatoria e (precision nomenclatural and
classificatoria classificatory)
immensità tranquilla e vuota (immensity calm and empty)
mondo frenetico e (world frenetic and congested)
congestionato
presenza ingombrante e vistosa (presence encumbering and
striking)
tempo continuo e immutabile (time continuous and
immutable)
- type AAS (ADJ+ADJ+N)
234
irreducibile sorda evidenza (irreducible deaf evidence)
bonari rozzi pupazzi (kind rough puppets)
monotone spennacchiate livree (monotonous plucked
plumages)
rosee variegate aiole (rosy variegated flowerbed)
stolidi scacazzanti piccioni (stupid craping pigeons)
ferma tonda faccia (firm round face)
- type AAA (ADJ+ADJ+ADJ) climax
maldestra, stonata, irresoluta (clumsy, tone-deaf, irresolute)
acre, greve, fradicio (acrid, heavy, rotten)
- type AeAeA (ADJ+CONJ+ADJ+CONJ+ADJ)
climax
grigia e opaca e arcigna (grey and opaque and sullen)
fulminea e duttile e prensile (as quick as lightning and ductile
and prehensile)
- type AoA (ADJ+DISJ+ADJ) correctio
aria malefica o insidiosa (appearance harmful or insidious)
onda contigua o successiva (wave contiguous or subsequent)
frasi allusive o sibilline (sentences allusive or sibylline)
aggettivo arcaico o vernacolo (adjective archaic or
vernacular)
References
Calvino I. 1983, Palomar, Einaudi, Torino.
Dardano M. 1978, La formazione delle parole
nell’Italiano di oggi, Bulzoni, Roma.
Mengaldo P.V. 1994, Il Novecento, il Mulino,
Bologna.
Mengaldo P.V. 1991, Aspetti della lingua di Calvino, in: La tradizione del Novecento, Terza
serie. Einaudi, Torino.
Picchi E. 1994, Statistical Tools for Corpus Analysis: a Tagger and Lemmatizer for Italian, in:
Martin W. et al (Eds), EURALEX94 Proceedings.
Roventini A. 1980, Concordanze delle Cosmicomiche. Analisi statistica del vocabolario e
delle variazioni stilistiche, Tesi di Laurea, Università di Pisa, Facoltà di Lettere e Filosofia.
Tekavcic P. 1980, Grammatica storica dell’italiano vol.III Lessico, il Mulino, Bologna.
Salva finestra- Tagger DBTPR7.PLO
--------------------------------------------------------------------------------------
--------
{Sost }altezza 1
{Sost }ampiezza 1
{Sost }bellezza 1
{Sost }certezza 2
{Sost }compattezza 1
{Sost }completezza 1
{Sost }compostezza 1
{Sost }debolezza 1
{Sost }fanciullezza 1
{Sost }fortezza 1
{Sost }franchezza 1
{Sost }freschezza 1
{Sost }giovinezza 1
{Sost }giustezza 2
{Sost }incertezza 4
{Sost }insicurezza 2
{Sost }leggerezza 2
{Sost }lucentezza 2
{Sost }lunghezza 1
{Sost }naturalezza 1
{Sost }piccolezza 1
{Sost }prontezza 2
{Sost }raffinatezza 2
{Sost }ricchezza 3
{Sost }rilassatezza 1
{Sost }riservatezza 1
{Sost }saggezza 2
{Sost }salvezza 1
{Sost }snellezza 1
{Sost }sostenutezza 1
{Sost }stranezza 1
{Sost }tristezza 1
- DBT -- E:Picchi---------------------------------------------------------------------
Fig. 1
235
D.B.T. (E.Picchi) Palomar
--------------------------------------------------------------------------------------
abbandonare V/S3IP
Forma : abbandona Freq. : 2
1)Copito de Nieve» non ** abbandona ** mai il copertone. Rif.: R077LRNN83X.45.Pag.0083.21
2)e si risponde non l’ ** abbandona ** mai: una variazione Rif.: R077LRNN83X.50.Pag.0119.26
2
abbattere V/S3IP
Forma : abbatte Freq. : 1
1) a che non s’ ** abbatte ** sulla riva: e dove Rif.: R077LRNN83X.140.Pag.0019.17
1
abitare V/S3IP
Forma : abita Freq. : 2
1) io egocentrico e megalomane che ** abita ** in lui. Ma Rif.: R077LRNN83X.25.Pag.0016.1
2), il geco che ** abita ** sotto le foglie su quel Rif.: R077LRNN83X.12.Pag.0058.15
2
abusare V/S3IP
Forma : abusa Freq. : 1
1) gli abusi di chi ** abusa **, egli non ha esitazioni Rif.: R077LRNN83X.116.Pag.0113.27
1
accentuare V/S3IP
Forma : accentua Freq. : 1
1) capta quasi altri dettagli e ** accentua ** l’ astrazione geometrica dell’ Rif.:
R077LRNN83X.60.Pag.0040.3
1
accettare V/S3IP
Forma : accetta Freq. : 1
1) presunta compagna, che lui ** accetta ** senza provarla. Rif.: R077LRNN83X.10.Pag.0101.13
1
acchiappare V/S3IP
Forma : acchiappa Freq. : 1
1) La trascura? No, ** acchiappa ** anche quella. La Rif.: R077LRNN83X.103.Pag.0061.14
1
.......
-------------------------------------------------------------------------------------------
Fig. 3
Salva finestra - Tagger DBTPR1.PLO
--------------------------------------------------------------------------------------
{Agg }acustico 2
{Agg }antico 1
{Agg }araldico 1
{Agg }arcaico 1
{Agg }aristocratico 1
{Agg }armonico 2
{Agg }asimmetrico 1
{Agg }astigmatico 1
{Agg }atavico 2
{Agg }biologico 1
{Agg }carico 2
{Agg }cosmico 1
{Agg }cronologico 1
{Agg }disarmonico 2
{Agg }egocentrico 1
{Agg }fisico 1
{Agg }frenetico 1
{Agg }genetico 1
{Agg }geometrico 1
{Agg }grafico 1
{Agg }identico 1
{Agg }logico 2
{Agg }meccanico 1
{Agg }microscopico 1
{Agg }neoplatonico 1
{Agg }pratico 1
{Agg }prospettico 1
{Agg }ritmico 1
{Agg }rustico 1
{Agg }scorbutico 1
{Agg }simbolico 1
{Agg }sintetico 1
{Agg }specifico 1
{Agg }storico 1
{Agg }unico 6
- DBT -- E.Picchi ------------------------------------------------------------------
Fig. 2
236
D.B.T. (E:Picchi) Palomar
--------------------------------------------------------------------------------------------
Andromeda W/
Forma : andromeda Freq. : 1
1) Nebulosa del Granchio o l’ addensarsi d’ un ammasso globulare in ** Andromeda ** non possono
non avere una qualche influenza sul funzionamento del Rif.: R077LRNN83X.52.Pag.0119.28
1
ArtNouveau W/
Forma : artnouveau Freq. : 1
1) ripiani di questa vetrina allineata una collezione di vasi _ ** ArtNouveau ** ó; la sera
una lampadina da 75 watt illumina gli Rif.: R077LRNN83X.8.Pag.0058.11
1
Arturo W/
Forma : arturo Freq. : 2
1) verde. Le stelle intorno sono tutte impallidite, tranne ** Arturo ** che brilla con aria
di sfida un po’ pi in alto Rif.: R077LRNN83X.21.Pag.0038.24
2) quasi ad accucciarsi sulle chiome degli alberi a nordovest; ** Arturo ** cala a picco sul
profilo della collina trascinando tutto l’ aquilone Rif.: R077LRNN83X.59.Pag.0046.3
2
Australia W/
Forma : australia Freq. : 2
1) le nuvole; una si stabilizza in forma e posizione d’ ** Australia **, e il signor Palomar
si convince che pi distinta vede Rif.: R077LRNN83X.40.Pag.0039.16
2) e il signor Palomar si convince che pi distinta vede quell’ ** Australia ** pi l’ obiettivo
a fuoco, ma nello stesso tempo Rif.: R077LRNN83X.42.Pag.0039.18
2
..........
-------------------------------------------------------------------------------------------
Fig.4