What we Want from GALLURA
Ephraim Nissan
and Yaakov HaCohen-Kerner
Dept. of Computing, Goldsmiths’ College, Univ. of London, 25–27 St. James, New Cross, SE14 6NW, London, U.K.
Department of Computer Science, Jerusalem College of Technology (Machon Lev), P.O.B. 16031, Jerusalem, Israel
Keywords: Information extraction, Explanation generation, Story generation.
Abstract: Information retrieval (IR) and, all the more so, knowledge discovery (KD), do not exist in isolation: it is
necessary to consider the architectural context in which they are invoked in order to fulfil given kinds of
tasks. This paper discusses a retrieval-intensive context of use, whose intended output is the generation of
narrative explanations in a non-bona-fide, entertainment mode subject to heavy intertextuality and strictly
constrained by culture-bound poetic conventions. The GALLURA project, now in the design phase, has a
multiagent architecture whose modules thoroughly require IR in order to solve specialist subtasks. By their
very nature, such subtasks are best subserved by efficient IR as well as mining capabilities within large
textual corpora, or networks of signifiers and lexical concepts, as well as databases of narrative themes,
motifs and tale types. The state of the art in AI, NLP, story-generation, computational humour, along with
IR and KD, as well as the lessons of the DARSHAN project in a domain closely related to GALLURA’s,
make the latter’s goals feasible in principle.
In the history of full-text IR, tools for retrieval from
very large historical corpora in Hebrew and Aramaic
were prominent, with the RESPONSA project (see
e.g. Choueka, 1989a, 1989b; Choueka et al. 1971,
Before the rise of Web search engines,
RESPONSA tools were the ones which achieved the
more far-reaching effects on society, because how
they empowered the retrieval of legal precedents in
rabbinic jurisprudence, thus affecting especially
legal practice of family law in Israel (as for family
law, in the Ottoman successor states, the usual
jurisdiction is the courts of the various religious
Religious cultures, as being the “consumers” of
religious texts, were, in a sense, the customers of a
considerable portion of early projects in IR: apart
from RESPONSA, whose corpora comprise the
Jewish texts from the sacred sphere through the
ages, this was also the case of Padre Busa’s Index
Thomisticus in Milan, and of the humanities
computing at the Abbey of Maredsous, in Belgium.
Exegesis (such as biblical interpretations) and
homiletics involve layers of texts, where a secondary
text refers to and either just quotes, or discusses,
some locus in the primary text; or then (as in the
Jewish aggadic midrash) expands on a biblical
narratives, filling the gaps where the primary text is
silent. Collections of aggadic midrash from late
antiquity (e.g., the Midrash Rabbah) or the Middle
Ages (e.g., Yalqut Shim‘oni) are a digest of a
multitude of homilies on biblical fragments of texts,
developing several often alternative ideas and
subnarratives. Cf. Hirshman (2006), Braude (1982),
Fishbane (1993), Hartman and Budick (1986).
* HyperJoseph is a hypertextual tool on the story
of Joseph in Genesis, with the secondary texts
elaborating on it (Nissan and Weiss, 1994).
* DARSHAN is a tool that invents homilies in
Hebrew (HaCohen-Kerner et al. 2007).
Retrieval in DARSHAN is intensive, and so is
the use of networks of lexical concepts.
DARSHAN generates ranked sets of either one-
sentence or one-paragraph homilies. While
producing its output, DARSHAN is able to quote
Nissan E. and HaCohen-Kerner Y..
DOI: 10.5220/0003688304790484
In Proceedings of the International Conference on Knowledge Discovery and Information Retrieval (KDIR-2011), pages 479-484
ISBN: 978-989-8425-79-9
2011 SCITEPRESS (Science and Technology Publications, Lda.)
from Scripture, to search for an occurrence
elsewhere in the textual canon, to replace words or
letters, to resort to puns, to interpret a word as an
acronym, and so forth. Use is made of patterns
which consist of canned text with places where to
plug in strings obtained through IR and
manipulation. The user supplies as input a biblical
verse, or a sentence, or a set of words, and also
specifies which devices should be applied. Filters
applied to the candidate output are alert, e.g., to
positive vs. negative connotation.
The quality of an individual output homily is
assessed as a sum of weighted factors, including:
length (as an indicator of complicacy); the
percentage of relevant words in the homily, out of
the total of words in the homily how many sentences
there are; how complex it was to insert every motif
into the homily generated; how many motifs were
actualized in the output homily being evaluated; how
many transformations were carried out; how many
words were replaced in the homily.
Having mentioned acronyms, consider that
HaCohen-Kerner et al. (2010b) discussed an
abbreviation disambiguation system for rabbinic
texts in Hebrew or Aramaic. Cf. Stock and
Strapparava (2005) on the HAHA project, whose
purpose is the humorous interpretation of acronyms.
As to connotations, Strapparava and Valitutti (2004)
described an affective extension of WordNet.
The GALLURA project seeks to develop software
that would interpret in Hebrew names by folk-
etymology, but in the context of a generated
narrative (aetiological tales, usually brief or even
very brief). The most closely studied model is a
large textual corpus of playfully creative writing that
embodies midrashic literary devices, by explaining
fancifully place-names of names for animal kinds.
The GALLURA project, now in the design
phase, requires, among the other things, capabilities
of story-generation, and of generating a playful
explanation. By themselves, these two tasks draw
upon three areas in AI:
explanation synthesis (for which, see e.g.
Schank, 1986,
1994; Walton, 2004),
story-generation (see e.g. Liu and Singh,
2002; Lönneker et al., 2005; and a long survey
in Nissan, 2011a: Ch. 5), and
computational humour (see e.g. Stock et al.,
2002; Ritchie, 2004; Waller et al., 2009).
Humour studies are interdisciplinary.
Moreover, GALLURA needs skills from
computational linguistics, including some that thus
far were modelled by linguistics, but not
folk-etymology (see e.g. Kirwin, 1985; Coates,
1994; Baldinger, 1973; Zuckermann, 2006),
phono-semantic matching (PSM), a discussion
of which is found in Zuckermann (2000, 2006).
For example, one of several PSM rules as
occurring in neologisation by adapting a foreign
term (Zuckermann 2000) is as follows (where SL is
the source language. TL is the target language):
SL y ‘b’ Æ TL
x ‘b’ Å TL x ‘a’
x is phonetically similar to y; a is similar to b
That is to say, the PSM introduced a new sense:
this was a PSM produced by shifting the meaning of
a pre-existent word in the target-language (TL).
Another rule of camouflaged borrowing (ibid.) is:
y ‘b’ÆTL
{x}+{z} ‘b’ÅTL {x} ‘a’, {z}
x is a lexical morpheme (e.g. root) that is
phonetically similar to y;
z is a grammatical morpheme (e.g. noun-pattern);
{x}+{z} is one word; a is similar to b
GALLURA should also have quality evaluation
capabilities, e.g., evaluating a story generated
(Peinado and Gervás, 2006), or evaluating morality
within a story (Reeves, 1991). We also need to
resort to computational argumentation: some such
current research into argumentation in computer
science looks into legal narratives (Bex, 2011).
Explanation as sought in GALLURA need not
necessarily be realistic; it is non-bona-fide (like in
humour), and must conform to a set of conventions,
of which realism is just a particular case (cf. Nissan,
2008). There are constraints on style: the output text
generated conforms to the early rabbinic linguistic
stratum and style (thus emulating the aggadic
midrash), with constraints on which lexical items or
morphological forms can be selected.
Rabbinic stylemes are the subject of current IR
including in the CUISINE text classifier.
So are the identification of rabbinic citations, and
chronological classification based on them. In fact,
HaCohen-Kerner et al. (2010a) discussed stylistic
feature sets for classification in CUISINE.
Automated identification of citations from rabbinic
texts has been researched (HaCohen-Kerner et al.,
2010c). Automated classification of rabbinic
KDIR 2011 - International Conference on Knowledge Discovery and Information Retrieval
PSM agent
Triggered upon Access)
Quotation Agent
Corpora & Tools)
Pool of
Motifs &
Emplotment a
responsa by period based on what they cite or are
cited by, was attempted successfully: HaCohen-
Kerner and Mughaz (2010) defined and effectively
applied “various kinds of ‘iron-clad’, heuristic and
greedy constraints defining the birth and death years
of an author based on citations referring to him or
mentioned by him.”
Several capabilities are required of GALLURA, and
many of them require retrieval. Fig. 1 shows
Coalition1 of agents, i.e. agents that often interact
among themselves. The control sequence is
opportunistic, according to the needs of the various
agents while they tackle a (sub) problem during a
particular run. They broadcast their need for help to
the other agents, and contract out the task. Some
agents however interact in a privileged manner with
one or more other gents, as they for a “coalition”.
Both the syntax agent and the stylemic agent
have to emulate early rabbinic language, but the pool
of stylemes and more abstract modes comprising
stylemes need actually be wider. Fig. 2 shows the
interplay of other coalitions of agents. In the
Lexicon, expected associations or behaviour are
triggered through demons, procedural code activated
upon access to individual lexical entries.
Coalition5 comprises an Encyclopedic agent, and
a Commonsense agent. The latter comprises two
modules: Concept-centred commonsense, and
Situational commonsense. Both the Emplotment
agent, and the Tex-generation agent closely interact
with the Argumentation agent.
It is usually proper nouns that are playfully
etymologised in the modern, archaising Hebrew
narrative corpus which is the main model for
GALLURA, and whose own model is the already
mentioned early rabbinic genre of the aggadic
midrash. Nevertheless, sometimes common nouns
are folk-etymologised as well, and most often these
are non-Hebrew words.
Here is a concise example. The input is Latin
aqua ‘water’. In the model corpus, there is this item:
Ma ra’ú Bnei Rómi, še-hém qorín et ha-
máyim ’aqwa (aqua)? Le-fí še-katúv:
“yiqqawú ha-máyim”.
Here is a translation of this Hebrew text:
Why [literally: what did they see], the
Romans [lit.: The Sons of Rome], that they
call water aqua? Because [lit.: to mouth of] it
is written [in Scripture]: “Let the water be
Figure 1: Coalition1 of agents.
Figure 2: The interplay of coalitions of agents.
In fact, the intertextual reference is to Genesis
1:9. The verbal form yiqqawú (passive future, 3rd
person plural) is from the root qwh. Corradicals one
can find in the Hebrew Bible include the verb and
noun for fluids gathering, for hoping and hope (the
word for ‘hope’ also has the little known sense
‘string’), and the noun now used for ‘line’.
Etymologically unrelated, Qwe also occurs, being
the name of a horse-trading land in Anatolia with
whom and with Egypt King Solomon traded in such
animals. Finding the apparent corradicals is trivial,
using the IR and NLP tools of the RESPONSA
project. What does require AI instead is for software
to be able to notice that Genesis 1:9, because it is
about water (and during an act of creation), is
splendidly apt an occurrence of the input aqua,
which PSM spuriously proposes as a derivative of
the root qwh (Semitic roots are “triliteral”).
There are features of the example considered,
pertaining to the lexicon, morphology, and style,
which clearly belong to the Mishnaic (i.e., early
rabbinic) historical stratum of Hebrew. Beginning
with a question, and in particular with one of the
many ways of asking ‘Why’ in Hebrew (i.e., lit.
“What did they see?”), which involve Coalition1 and
Coalition2, the Lexicon.
Asking and answering here also involves some
rather rudimentary involvement of argumentation. A
shortcut would be to use a canned-text encoding of a
pattern, in the manner of DARSHAN. Actually
however there is some sophistication in the example
considered, because we are not abstractly taking
about Latin; rather, the expression is made concrete,
with the Sons of Rome being invoked from the Pool
of stock characters. This dovetails with the
underscoring of their agency, when the option
selected for saying ‘Why’ is “What did they see?”
The following would be a much more difficult
example for GALLURA to replicate, and both
retrieval and manipulation would be intensive and
laborious. In the model corpus we use, place-names
around the world are explained by both playful
etymology, and fantasy history narratives. It is often
the case that a story is told about one of the human
groups leaving the Tower of Babel. The Generation
of the Division (Dór ha-Pallagá) or the Ones
Leaving the Tower (Yots’ei ha-Migdal) would be
often resorted to in GALLURA’s Pool of stock
characters. Let us consider a story on Laos.
“Teach us, Sir” (yelammédenu Mar, a cliche
especially associated with the lost rabbinic Midrash
Yelammedenu), “What did the Nations see” (i.e.,
‘why’: má ra’ú ha-’ummót), “that they call” (še-
qorín: a Mishnaic verbal inflection) “one of them
Laos” (achát mehén Lá’os). “I shall answer you
immediately!” (Af aní mešivkhem mi-yád! a cliché).
A ready pattern of argumentation: “Instead of [lit.:
Until] you asking why that nation is called Laos”
(‘Ád še-attém šo’alín lámma otáh ’ummá qruyá
Lá’os), “be asking what did the Sons of Greece see”
(hevú šo’alín ma ra’ú Bnei Yaván), “that all
populations” (še-kól ’okhlosín, itself a Green
loanword in Hebrew) “were called in their mouths
[i.e., by them] λαός” (niqre’ú be-fihém lá’os). “Once
the Ones Leaving the Tower went out of Babel”
(Keván še-yats’ú Yots’éi ha-Migdál mi-Bavél), “they
were tired (le’ín) and exhausted on the road” (hayú
le’ín u-me‘uyyafín ba-dárekh).
Sustained walking is tiresome, and one term for
‘tired’ is related by PSM to Laos. Now, consider that
in a crowd (a spawned demon would inform
GALLURA), you would expect somebody trying to
sell snacks and drinks, unless circumstances exclude
this (e.g., if it’s a day of fast, or a famine causes
starvation). Such a situational cliché is funny if it
does not quite match the situation at hand. The
theme of the exodus from Babel, in the model
corpus, often has a wise old man advise the crowd,
but some other time, some individual takes
advantage, being cunning rather than altruistic.
“The more astute among them” (‘Armumiyyín se-
bahém), “who were traders and vendors of edibles”
(še-hayú ba‘aléi praqmátya [a typical early rabbinic
term] u-mokhréi mezonót), “this way they were
speaking to them” (kha hayú ’omrím lahém): “Let
the legs be strong!” (Techezáqna ha-ragláyim!). The
latter contains a Biblical Hebrew verbal form, the
3rd person plural feminine (as ‘legs’ are feminine in
Hebrew), whereas Mishnaic Hebrew discarded that
form, using the masculine. As this is a modified
quotation, using a Biblical Hebrew morphological
(or lexical) form is legitimate for GALLURA. “Let
the legs be strong!” (Techezáqna ha-ragláyim!) is a
modification of “Let the hands be strong!”
(Techezáqna ha-yadáyim!), the title of a famous
labour song by Bialik. Such a temporal flashforward
for a story set at the times of the Tower of Babel is a
funny transgression (rather than an insipient
“Whatever you shall put under your teeth, you
shall find in your legs!” (Má še-tittnú táchat
šinneikhém, timtse’ú be-ragleikhúm!). This is a
Hebrew adaptation of an Aramaic early rabbinic
proverb. “Be chewing” (hevú lo‘asín, associated by
PSM with Laos), “as for this you were created!” (še-
KDIR 2011 - International Conference on Knowledge Discovery and Information Retrieval
la-zé notsártem! This is evocative of le-khakh
notsarta, “for that purpose [of studying] thou hast
been created”, in Maxims of the Fathers, 2:8).
“Every population” (lit.: population population,
’okhlosín ’okhlosín), “all of them are chewers!”
(kullam la‘osot!). “As they were hearing them
saying so” (Keván še-hayú som‘ín ’otám ‘omrím
ken, partly a quotation of how the crowd in the
Temple used to respond to a given utterance of the
High Priest on the Day of Atonement, a day of fast),
“their saliva flowed, they paid the price, would take
and eat” (záv hayá rirám, notnín mamón, notlín ve-
‘okhlín, with typical Mishnaic wording).
Clearly, obtaining from GALLURA output such
as this story from our model literary corpus would
be as ambitious a goal as it can get. Anything in the
middle would be nice to achieve. See Nissan’s
(2011b) 150-page discussion of playful narrative
It is important to realise that knowledge discovery or
information extraction as involved in accessing the
historical textual canon as well as ontologies and
representations of commonsense or encyclopedic
knowledge, can be easier in one direction, while
very difficult in the other. We exemplify this with an
item from our model corpus. Edom (medieval for
‘Europeans’) in the Land of Ashkenaz (medieval for
‘Germany’) the text relates, for many generations
were eager to insert fon (i.e., von) before their family
names, as it would signal their patrician ancestry.
“What did cause that? The episode of Eldad and
Medad caused that”, at Numbers 11:26–29. Moses
appointed seventy elders, but those two did not
come, and prophesised nevertheless. Joshua tells
Moses to put them under arrest, but Moses retorts:
“Are you jealous on my behalf? If only” all the
people were prophets.
U-mí yittén (lit.: “And who would give”) was
rendered, in the canonical Jewish Aramaic
translation (the Targum by Onqelos) as: Ra‘ena fon,
i.e., “I wish fon”, where fon (a grammaticalised
denominal conjunction) means any of ‘face’, ‘turn’,
‘that would’, or ‘lest’. If you were reading Onqelos,
you may happen to notice this locus serendipitously.
But had you begun with eagerness for ennoblement,
it would be very difficult to devise an appropriate
search that would retrieve a biblical “I want fon”.
GALLURA is an ambitious project, now in the
design phase, requiring the interplay of various
agents or coalitions of agents specialised per domain
of expertise. Several of these agents have retrieval-
intensive requirements. GALLURA has to devise
playful etymologies with a backup story to go with.
It builds upon the experience and part of the
architectural features of DARSHAN — especially
how the pool of devices is organised, and the
approach to retrieval, which is mostly from the same
textual corpora. GALLURA is much more difficult
to achieve, but at the stage reached by a number of
domains within AI, NLP, IR, and KD, it is in
principle feasible. Any progress on any part of the
architecture would by itself be a valuable
achievement. A global advantage already at present,
in this project, is that thanks to manual analysis of
many items in the creative writing corpus which is
our main model, it is possible to model
algorithmically all devices required.
Baldinger, K. 1973. À propos de l’influence de la langue
sur la pensée: Étymologie populaire et changement
sémantique parallèle. Revue de Linguistique Romane,
37, pp. 241–273.
Bex, F. 2011. Arguments, Stories and Criminal Evidence:
A Formal Hybrid Theory, Law and Philosophy Series,
vol. 92. Springer, Dordrecht.
Braude, W. G. 1982. Midrash as deep peshat. In: S. R.
Brunswick (ed.), Studies in Judaica, Karaitica and
Islamica (Presented to Leon Nemoy on his Eighties
Birthday). Bar-Ilan University Press, Ramat-Gan,
Israel, pp. 31–38 [English].
Choueka, Y. 1989a. RESPONSA: An operational full-
text retrieval system with linguistic components for
large corpora. In: E.I. Cuomo (ed.), Law in
Multicultural Societies, Proceedings of IALL, the
International Association of Law Libraries Meeting,
Jerusalem, 1985. The Hebrew University, Jerusalem,
1989, pp. 47–82.
Choueka, Y. 1989b. Responsa: A full-text retrieval system
with linguistic processing for a 65 million-word
corpus of Jewish heritage in Hebrew. In a Special
Issue on non-English Interfaces to Databases, IEEE
Data Engineering, 12(4), pp. 22–31.
Choueka, Y., Cohen, M., Dueck, J., Fraenkel, A.S., Slae,
M. 1971. Full-text Document Retrieval: Hebrew
Legal Texts (Report on the first phase of the Responsa
Retrieval Project). In: M. Minker, S. Rosenfeld (eds.),
Proceedings of the ACM Symposium on Information
Storage and Retrieval, Maryland, 1971. Association
for Computing Machinery, New York, 1971, 61–79.
Choueka, Y., Fraenkel, A. S., Klein, S.T., Segal, E. 1987.
Improved techniques for processing queries in full-text
systems. In: C.T. Yu, C.J. van Rijsbergen (eds.),
Proceedings of the Tenth Annual International ACM-
SIGIR Conference on Research and Development in
Information Retrieval, New Orleans 1987. ACM, New
York, 1987, pp. 306–315.
Coates, R. 1994. Folk etymology. In: R.E. Asher (ed.),
The Encyclopedia of Language and Linguistics,
Pergamon Press, Oxford, Vol. 3, pp. 1267–1270.
Fishbane, M., ed. 1993. The Midrashic Imagination,
University of New York Press, New York.
HaCohen-Kerner, Y., Mughaz, D. 2010. Estimating the
birth and death years of authors of undated documents
using undated citations. Proceedings of the Seventh
International Conference on Natural Language
Processing (IceTAL 2010), August 16–18, 2010,
Reykjavik, Iceland (LNCS 6233), pp. 138–149.
Springer-Verlag, Berlin.
HaCohen-Kerner, Y., Avigezer, T.S.-T., Ivgi, H. 2007.
The Computerized Preacher: A prototype of an
automatic system that creates a short rabbinic homily
[Hebrew]. B.D.D. (Bekhol Derakhekha Daehu):
Journal of Torah and Scholarship (Bar-Ilan
University, Ramat-Gan) 18, pp. 23–46.
HaCohen-Kerner, Y., Beck, H., Yehudai, E., Rosenstein,
M., Mughaz, D. 2010a. Cuisine: Classification using
stylistic feature sets and/or name-based feature sets.
Journal of the American Society for Information
Science and Technology, 61(8), pp. 1644–1657.
HaCohen-Kerner, Y., Kass, A., Peretz, A. 2010b. A
Hebrew Aramaic abbreviation disambiguation system.
Journal of the American Society for Information
Science and Technology, 61(9), pp. 1923–1932.
HaCohen-Kerner, Y., Schweitzer, N., Shoham, Y. 2010c.
Automatic identification of biblical quotations in
Hebrew-Aramaic documents. Int. Conf. on Knowledge
Discovery and Information Retrieval (KDIR), pp. 320–
325, Oct. 2010, Valencia.
Hartman, G. H., Budick, S., eds. 1986. Midrash and
Literature, Yale University Press, New Haven, CT.
Hirshman, M. 2006 . Aggadic midrash. Ch. 2 in: S.
Safrai, Z. Safrai, J. Schwartz, P. J. Tomson (eds.), The
Literature of the Sages, Second Part, Royal Van
Gorcum, Assen, Netherlands, and Augsburg Fortress
Press, Minneapolis, MN, pp. 107–132.
Kirwin, W. 1985. Folk etymology: Remarks on linguistic
solving and who does it. Lore and Language
(Sheffield, U.K.), 4(1), pp. 18–24.
Liu, H., Singh, P. 2002. MAKEBELIEVE: Using
commonsense knowledge to generate stories. In Proc.
of the 18th National Conf. on Artificial Intelligence
and 14th Conf. on Innovative Applications of Artificial
Intelligence, pp. 957–958.
Lönneker, B., Meister, J. C., Gervás, P., Peinado, F.,
Mateas, M. 2005. Story generators: Models and
approaches for the generation of literary artefacts. In
the ACH/ALLC-2005 Conference Abstracts, Victoria,
BC, Canada, June 15–18, 2005, pp. 126–133.
Nissan, E. 2008. Chance vs. causality, and a taxonomy of
explanations. In: M. Negrotti (ed.), Natural Chance,
Artificial Chance, thematic volume of Yearbook of the
Artificial, Vol. 5. Peter Lang, Basel, pp. 195–258.
Nissan, E. 2011a. Computer Applications for Handling
Legal Evidence, Police Investigation, and Case
Argumentation. Springer, Dordrecht.
Nissan, E. 2011b. A Study of Humorous Explanatory
Tales. In: N. Dershowitz and E. Nissan (ed.),
Language, Culture, Computation: Essays in Honour of
Yaacov Choueka. Springer, Berlin, in press.
Nissan, E., Weiss, H. 1994. The HyperJoseph project (2
parts). In: F. Poswick (ed.), Proc. 4th International
Conference on Bible and Computers (AIBI’94),
Amsterdam, August 15–18, 1994. Champion-Slatkine,
Geneva & Paris, 1995, pp. 154–162 & 163–173.
Peinado, F., Gervás, P. 2006. Evaluation of automatic
generation of basic stories. In a special issue on
Computational Creativity, New Generation
Computing, 24(3), pp. 289–302.
Reeves, J. 1991. Computational Morality: A Process
Model of Belief Conflict and Resolution for Story
Understanding, Tech. Rep. 910017, Comp. Science
Dept.. Univ. of California, Los Angeles. ftp: // ftp.
Ritchie, G. 2004. The Linguistic Analysis of Jokes,
Routledge, London.
Schank, R. C., ed. 1986. Explanation Patterns:
Understanding Mechanically and Creatively,
Lawrence Erlbaum Associates, Hillsdale, NJ.
Schank, R. C., Kass, A., Riesbeck, C. K., eds. 1994.
Inside Case-Based Explanation, Erlbaum, Hillsdale, NJ.
Stock, O., Strapparava, C. 2005. The act of creating
humorous acronyms. Applied Artificial Intelligence,
19(2), pp. 131–151.
Stock, O., Strapparava, C., Nijholt, A., eds. 2002. The
April Fools’ Day Workshop on Computational
Humour: Proceedings of the 20th Twente Workshop
on Language Technology (TWLT20), Trento, Italy,
April 2002. University of Twente, The Netherlands.
Strapparava, C., Valitutti, A. 2004. WordNet-Affect: An
affective extension of WordNet. Proceedings of the
Fourth International Conference on Language
Resources and Evaluation (LREC 2004), Lisbon, May
2004, pp. 1083–1086.
Waller, A., Black, R., O’Mara, D. A., Pain, H., Ritchie,
G., Manurung, R. 2009. Evaluating the STANDUP
pun generating software with children with cerebral
palsy. ACM Transactions on Accessible Computing
(TACCESS), 1(3), article no. 16, at the ACM site.
Walton, D. N. 2004. Abductive Reasoning, University of
Alabama Press, Tuscaloosa, Alabama.
Zuckermann, G. 2000. Camouflaged Borrowing: Folk-
Etymological Nativization in the Service of Puristic
Language Engineering, D.Phil. Dissertation in
Modern Languages, University of Oxford, Oxford.
Zuckermann, G. 2006. “Etymythological othering” and the
power of “lexical engineering”. Ch. 16 in T. Omoniyi,
J.A. Fishman (eds.), Explorations in the Sociology of
Language and Religion, Benjamins, Amsterdam, pp.
KDIR 2011 - International Conference on Knowledge Discovery and Information Retrieval