=Paper=
{{Paper
|id=None
|storemode=property
|title=Exploring Semantic Orientation of Adverbs
|pdfUrl=https://ceur-ws.org/Vol-757/paper_8.pdf
|volume=Vol-757
}}
==Exploring Semantic Orientation of Adverbs==
https://ceur-ws.org/Vol-757/paper_8.pdf
Exploring Semantic Orientation of Adverbs
Potemkin, S.B., Kedrova, G.E.
Philological faculty, Moscow State University, Moscow, Russia
{potemkin,kedr}@philol.msu.ru
Abstract. Sentiment analysis often relies on a semantic orientation lexicon of
positive and negative words. Determining the semantic orientation of words is
necessary for correct estimation of the content of statements in the media, In-
ternet, in the writings and speech. Qualitative adverbs expressing evaluation, in-
tensity, direction of action are important as the modifiers of the main sentence
predicate. In this paper we propose a method for extracting seed set of adverbs
from a collection of pairs of antonym. A model based on the representation of a
set of synonyms from the Russian lexicons as a graph, and determination the
semantic orientation of the adverbs concerning three main dimensions of the
semantic differential also demonstrated. The assessment of performance of the
method in comparison with the dictionary data shows effectiveness of the
method obtained.
Keywords. Sentiment, semantic differential, antonyms, seed set, graph.
1 Introduction
Nowadays, the availability of resources for Natural Language Processing (NLP) re-
mains a hot topic, in particular for Russian especially due to the lack of comprehen-
sive semantic resources, despite efforts made to provide a freely-available Russian
WordNet [1]. Ability to establish relativity, similarity, or semantic distance between
words and concepts is the basis of computational linguistics. This paper deals with
measuring of distance within the syntactic category of adverbs. This set of words is
crucial for some applications because adverbs modify or clarify the meaning of other
words (verbs, nouns, adjectives). The adverbs are of particular interest to determine
the semantic orientation of syntagma containing a main word and its modifier (ad-
verb). Measuring the semantic distance or similarity between the English words most
often is based on WordNet [2], and almost exclusively on taxonomic relationships
established in this database. So such approach is applicable only to the syntactic cate-
gories of nouns and verbs.
The aim of this paper is to extract a list of semantically oriented adverbs and de-
velop the measure of proximity based on dictionaries of synonyms. The article is
structured as follows. In Section 1 the problem of extracting the seed set of semanti-
cally oriented adverbs from the lexicon of Russian antonyms is discussed. In Section
2 we describe the previously proposed measures of semantic distance between words,
as well as an elementary way to map synonyms onto a graph. In Section 3 the basic
characteristics of the subjective understanding of the meaning and the measures based
on the distance in a graph of synonyms are discussed. Finally, Section 4 presents
some results and conclusions. Additionally, we explore the use of visualization tech-
niques to gain insight into the results obtained.
2 Extracting the seed set of adverbs
A number of approaches have been proposed for creating semantic orientation
lexicons in English, most of them are computationally expensive and rely on signifi-
cant manual annotation and large corpora. Particularly, the General Inquirer [3] cre-
ated in the beginning of the last century is used as the gold standard for assessment
the quality of new-generated lexicons. For Russian language there is no open-source
and reliable lexicon with positively and negatively marked entries. We propose some
approaches to generate a broad coverage semantic orientation lexicon for Russian
adverbs which includes both individual words and multi-word adverbial expressions
using only dictionaries of antonyms and synonyms, requiring a small amount of man-
ual pruning and database processing.
First of all we have analyzed a list of antonyms collected from published dictionar-
ies of antonyms [4,5]. This list contains 7300+ antonymous pairs (adjectives, nouns,
verbs, adverbs and prepositions as well). The semantically oriented words were ma-
nually extracted from this list and arranged in 2 separate lists – positive (1859) and
negative (2229) words. This seed lexicon could be compared with the GI lexicon
which contains orientation labels for only about 3600 entries.
Next step was to extend our seed lexicon to obtain a broad coverage of different
texts under consideration concerning sentiment analysis. Automatic approaches to
create (English) semantic orientation lexicon and, more generally, approaches for
word-level sentiment annotation can be grouped into two kinds: (1) those that rely on
manually created lexical resources—most of which use WordNet; and (2) those that
rely on text corpora [6]. As a lexical source we use a structured list of Russian syno-
nyms collected from a number of published and Internet-available dictionaries such as
[7] and others (11 sources). List of synonyms contains ~600000 word-pairs including
~10000 pairs of adverbs. All synonyms {s(wi)} of each seed word wi receives the
same semantic orientation as wi. The number N of occurrences of a synonym s(wi) in
the extended set contributed by different seed-words wi, (i=1…N) indicates the confi-
dence of semantic orientation. After manual pruning we have got a list of positively
marked (5990, including 731 adverbs) and negatively marked (6853, including 592
adverbs) words. Since the most part of Russian adverbs could be derived as the short
form singular neutral or short form plural adjective (3135) the list of semantically
orientated adverbs could be expanded.
3 Measures of distance
A number of distance or similarity measures exist for English based (completely or
partially) on WordNet. In particular, such measure is defined as the number of edges
of the path through the taxonomic relations (IS-A, Part-of, or WordNet’s hyponymy
relation). In [8] the concept of bond length was extended for all relations in WordNet
by their clustering in the horizontal (synonyms) or vertical (hyponymy) direction and
assigning a penalty for changing the direction of the path motion. Overview of five
measures and evaluation of their effectiveness using the associations between the
words is given in [9]. Exclusive usage of hyponymy delimits the measure of distance
or similarity only to the syntactic categories of nouns and verbs, as hyponymy rela-
tions in WordNet are established only for these grammatical categories. Therefore,
such measures could not be applied to adjectives and adverbs.
The semantic distance between the words could be determined in the similar way
as the definition adopted in graph theory [10]. The simplest approach is just to gather
all the words from the Dictionary of synonyms and to link each member of a synony-
mous group with its dominant word as indicated in the Dictionary. Let G(W,S) be the
undirected graph, with W the set of nodes being all the words from the Dictionary
with associated part-of-speech, S - the set of edges connecting each member of syn-
onymous group with its dominant word. Every group of synonymous words could be
connected to each other and form a clique in G graph. A path P is the sequence of
nodes connected by edges of G and geodesic is the shortest path between two nodes.
Geodesic distance, D(wi,wj) between two words wi and wj is the length (number of
edges) of the shortest path between wi and wj. If there is no path between wi and wj,
the distance between them is infinity. The minimal path-length defines a metric on the
set of synonyms. All axioms of the metric space are fulfilled in this case. Usually
synonymous groups comprises the words of the same grammatical category and entire
graph G is decomposed into disjoint sub-graphs or networks for nouns, verbs, adjec-
tives and adverbs. (Fig. 1). In each network exists a maximal connected component
that contains 70-90% of all nodes of the graph constructed from the Dictionary of
synonyms. Maximum component in the class of Russian adverbs contains about 8500
words. The words in this connected component could be analyzed using the metric
defined by the length of geodesics.
4 Semantic orientation of adverbs
Classical work on the measurement of emotional or affective values in texts is the
theory of semantic differential by Charles Osgood. Word meaning in cognitive psy-
chology, is “a strictly psychological one: those cognitive states of human language
users which are necessary antecedent conditions for selective encoding of lexical
signs and necessary subsequent conditions in selective decoding of signs in mes-
sages.” [11]. Semantic differential method was applied mainly to the adjectives meas-
ured in such dimensions as active/passive, good/bad, positive/negative, beautiful/ugly,
etc. Each pair of bipolar adjectives is a factor or an axis in the method of semantic
differential. Application of factor analysis to extensive empirical material gave an
unexpected result: most of the variance in judgment could be explained by only three
major factors including the evaluative factor (e.g., positive/negative); the potency
factor (e.g., strong/weak); and the activity factor (e.g., active/passive). Among these
three factors, the evaluative factor has the strongest relative weight for determining
the semantic orientation.
Turning to the selected Russian adverbs, we note that the vast majority of adverbs
is matched with the words of other parts of speech primarily with the adjectives
(cheerful – cheerfully // бодрый - бодро, brutal – brutally // жестокий - жестоко), so
that the semantic differential can be naturally extended to motivated adverbs, which
bear semantic meaning and, accordingly, deliver the information on their semantic
orientation. All three pairs of bipolar adverbs negatively/positively (плохо/хорошо);
weakly/strongly (слабо/сильно), passively/actively (пассивно/активно) are con-
tained in the maximal component of the sub-graph of synonymous adverbs Gadv. One
can assume that the distance to positively (хорошо) is a measure of positive assess-
ment of an adverb. However, it is easy to show that this measure is in fact rather con-
troversial.
Fig. 1. A fragment of the maximum connected component subgraph of adverbs, Gadv.
A striking example of this is that the words positively (хорошо) and negatively
(плохо) are closely related through the path of synonyms. There is a sequence of only
5 words in English (negatively, hardly, tightly, thoroughly, comprehensively, soundly,
positively), and 6 words in Russian (плохо, дешево, легко, просто, совсем, очень,
здорово, хорошо) – see Fig. 1 – connecting opposites, each pair of words in this se-
quence is certainly synonymous (at least in one of their meanings). Thus, we find that
d(positively, negatively) = 6; d(хорошо, плохо) = 7. Despite the fact that the adverb
positively (хорошо) and negatively (плохо) have opposite meanings, they are closely
related by synonymy path. Of course, this is not due to any error in the Dictionary of
synonyms. Partial explanation lies in the wide use of two Russian adverbs хорошо
(625 ipm), плохо (187 ipm) [12]. Due to the fact that both words are members of the
maximum connected component of Gadv sub-graph, we can consider not only the
shortest distance from any adverb to "positively", but the shortest distance to its anto-
nym, "negatively". This idea is concretized [13] in the definition of EVA function,
which allows to measure the relative distance from the word of two opposites, "posi-
tively" and "negatively":
EVA (w) = (d (w, neg)-d (w, pos)) / d (neg, pos).
Under the assumption that there is no word "worse than negatively" or "better than
positively" the values of EVA lie in the interval [-1,1], for example, the word "hon-
estly" is evaluated by function EVA (honestly) gives a value of 1 as follows EVA
(honestly) = (d (honestly, neg) - d (honestly, pos)) / d (pos, neg) = (8-2) / 6 = 1,. The
measures for other Osgood’s dimensions is defined similarly. For the potency factor
the function: POT (w) = (d (w, weakly)-d (w, strongly)) / d (strongly, weakly) is de-
fined; for the activity factor the function: ACT (w) = (d (w, passively)-d (w, actively))
/ d (actively, passively) is defined. This fact allows to define measures for any two
words belonging to the maximal connected component of the adverbs subgraph.
An assumption on the boundary position of words negatively/positively is not en-
tirely justified. Intuitively, perfectly (превосходно) is better than positively, disgust-
ingly (отвратительно) is worse than negatively. Bearing this in mind and using the
geometry of a triangle with vertices {w, pos, neg}, we redefine the function of EVA,
namely:
EVA1(w)=(d(w, neg)-d(w, pos))*(d(w, neg)+d(w, pos))/d2(neg, pos).
The values of EVA1 sometimes are beyond the interval [-1,1]. Similarly, we can
redesign POT(w) and ACT(w).
For English adjectives (and motivated adverbs) there exists the source for assessing
the measure constructed above in comparison with the independently obtained an-
swers to the "General Inquirer” [11], which contains a set of words to assess three
Osgood’s factors. Word lists were obtained from the Stanford political dictionary,
where each of the 3000 most frequent common words were assessed by three or more
experts concerning each Osgood’s factor. Thus 765 positive and 873 negative words
for the assessment factor were obtained, 1474 strong and 647 weak word for the po-
tency factor and 1568 active and 732 passive words for activity factor. Comparison of
results obtained with the General Inquirer gave the values of 70 - 80% of matches,
depending on what words were considered as neutral in terms of EVA function.
In the absence of available data for content analysis we used the Russian dictionar-
ies of antonyms as an independent source. Antonymous pair is a pair of words (or
rather, the specific meanings of words), one opposed to the other on semantic
grounds, such as hot - cold fast - slow, present - absent. We suggest that adverbs be-
longing to pair of antonyms lie on the "opposite sides" of the entire set of adverbs.
Methods of multidimensional scaling deliver a mapping of multidimensional space
with the defined distance between individual points d(wi, wj) onto a space of smaller
dimension, namely the plane (Fig. 2). Figure 2 a, b shows that the pairs of antonyms
lie near the diameters of the set of adverbs. For a more profound study of the structure
of the space of adverbs we have constructed chains of synonyms connecting anto-
nyms pairs within the sub-graph Gadv.
Fig. 2. Two chains of synonyms, joining antonymous pairs of adverbs.
a) up – a consistent path; b) down – an inconsistent result.
Chain in Fig. 2a is a consistent result, i.e. the chain of synonyms passes on the pe-
riphery of the set of adverbs and the distances between the synonyms do not exceed
the distance between the antonyms. Unfortunately, the situation is not always as fa-
vorable. In Fig. 2b pair of antonyms is close to the diameter, but the chain of syno-
nyms is not at the periphery of the set, but lays in the central part of the set, alternates
its direction, and the distances between synonyms is often greater than the distance
between antonyms. Probably it is necessary to determine more accurate distance be-
tween the words and to choose correctly the axes of the adverb space using the princi-
pal components method. These new axes should not coincide Osgood’s dimensions.
5 Discussion and conclusions
In this paper we define a measure of the distance between adverbs using synonyms
graph. It seems obvious that the choice of similarity measure, or distance largely de-
pends on the type of the problem. The choice of distance measure on the grounds of
synonyms is connected with the goal of determining the semantic orientation of ad-
verbs. In contrast to Osgood’s semantic differential associated with the reaction of
people on the stimulus - words presented, or the possible emotional impact of words,
this model is based solely on the lexical material and is intended to represent rela-
tively objective meanings which are fixed in Dictionaries. Further studies will deter-
mine the semantic orientation of sentences or the whole text on the basis of the orien-
tation of its constituent words. Our method allows to evaluate other classes of words
such as nouns, adjectives and verbs, but this extension will require a significant in-
crease of calculations and special methods for processing large data sets, since an
algorithm for computing shortest paths requires O(n3) operations, where n is the num-
ber of words in graph G(W,S).
References
1. Azarova, I.V., Sinopalnikova, A.A., Javorsky, M.V. Principles of construction of WordNet-
thesaurus RussNet (in Russian) In: Computer linguistics and intellectual technologies. -
Proceedings of the International conference Dialogue'2004 pp.542−547.Мoscow (2004)
2. Fellbaum, C. (ed.), WordNet: An Electronic Lexical Database // Language, Speech, and
Communication Series. The MIT Press, Cambridge MA (1998)
3. Stone, P.J. Thematic text analysis: new agendas for analyzing text content. In: C. Roberts
(ed.), Text Analysis for the Social Sciences. Lawrence Erlbaum Associates, Mahwah NJ
(1997).
4. Vvedenskaia, L.A. Dictionary of Russian Antonyms (in Russian) OOO «AST Publishing
House », Moscow (2004)
5. Lvov, M.R. Dictionary of Russian Antonyms (in Russian) “Press-Kniga” (2006)
6. Mohammad, S et.al Generating High-Coverage Semantic Orientation Lexicons From
Overtly Marked Words and a Thesaurus In: Proceedings of the Conference on Empirical
Methods in Natural Language Processing (EMNLP), Singapore (2009)
7. Alexandrova, Z.E. Dictionary of Russian Synonyms (in Russian) “Russkii Iazyk-Media,
Moscow (2005)
8. Hirst, G., St-Onge, D. Lexical Chains Representations of Context for the Detection and
Correction of Malapropisms". In Fellbaum, C. (ed.) WordNet. An Electronic Lexical Da-
tabase, The MIT Press, (1998).
9. Budanitsky, A., Hirst, G. Semantic distance in WordNet: An experimental, application-
oriented evaluation of five measures. In: Workshop on WordNet and Other Lexical Re-
sources. Second meeting of the NAACL, Pittsburgh, (2001)
10. Potemkin, S.B. Semantic Distance over Lexical Database and WordNet In: Proceedings of
10 International Conference «Cognitive Modeling in Linguistics », CML, Montenegro Be-
chichi (2008).
11. Osgood, C.E., Succi, G.J., Tannenbaum P.H., The Measurement of Meaning.. University
of Illinois Press, Urbana IL (1957).
12. Sharov, S. Frequency Dictionary of Russian
http://www.artint.ru/projects/frqlist.asp (2003)
13. Kamps, J., Marx, M., Robert, J., Mokken, M. Using WordNet to Measure Semantic Ori-
entations of Adjectives In: Proceedings of the 4th International Conference on Language
Resources and Evaluation (LREC'04, Vol. IV ) pp. 1115-1118. (2004)