A New Data Structure for Processing Natural Language Database
Queries
Richard A. Frost
a
and Shane Peelar
b
School of Computer Science, University of Windsor, 401 Sunset Avenue, Windsor, Ontario, Canada
Keywords: Natural Language Processing, Natural Language Query Interfaces, Compositional Semantics, Event
Semantics, Quantification.
Abstract:
Natural Language Query Interfaces (NLQIs) have once again captured the public imagination, but developing
them has proven to be non-trivial. One way is by using a Compositional Semantics (CS) to directly compute
the answer to a query from the meanings of its parts. The query is treated as an expression of a formal
language, and interpreted directly with respect to a database which provides meanings for words, which are
the basic components. The meanings of compound phrases in the query, and the answer to the query itself,
are computed from their constituent words and phrases using semantic rules that are applied according to the
query’s syntactic structure. Montague Semantics (MS), which is a type of CS, has been used in various NLQIs
previously. MS accommodates common and proper nouns, adjectives, conjunction and disjunction, intransitive
and binary transitive verbs, quantifiers, and intentional and modal constructs. MS does not provide an explicit
denotation for n-ary transitive verbs nor does it provide an explanation of how to handle prepositional phrases.
By adding events to MS and by introducing a new data structure, transitive verbs and prepositional phrases
can be accommodated as well as other NL features that are often considered to be non-compositional.
1 INTRODUCTION
We begin by describing a Natural Language Query In-
terface (NLQI) that we have built. We hope that the
interface will motivate readers to look into our mod-
ifications to MS. In Section 2, we explain how our
NL Query interface (NLQI) can be accessed through
the Web. In Sub-section Section 2.1, we describe
the Semantic Web triplestore. In Section 3 we dis-
cuss example queries and their results: in Section 3.2,
we provide examples of what are often referred to as
“non-compositional” features of NL that our NLQI
can handle, and in Section 3.3 we give examples of
NL structures that could be accommodated by exten-
sions to our approach. With each of the examples we
provide an informal explanation of how the answer is,
or could be, computed.
In Section 4, we describe the new FDBR data
structure which is central to our approach, and which
can be created from an event-based triplestore (as we
do in our online NLQI), or from a relational database.
Much of our semantics is based on MS. We differ
a
https://orcid.org/0000-0001-7083-5060
b
https://orcid.org/0000-0001-7391-0951
in these ways:
1. We add events to the basic ontological concepts of
entities and truth values.
2. Each event has a number of roles associated with
it. Each role has an entity as a value.
3. For efficiency, we use sets of entities rather than
characteristic functions of those sets as is the case
in MS.
4. We define transitive n-ary verbs in terms of sets of
events, each with n roles.
5. We compute FDBRs, the novel datastructure pre-
sented in this paper, from sets of events (could be
computed from relations), and use them in the de-
notations of transitive verbs, and in computing re-
sults of queries containing prepositional phrases.
Although not referred to as an FDBR, the use of
relational images in denotations of verbs was first
proposed by (Frost and Launchbury, 1989).
We hope that this paper reawakens an interest in
Compositional Semantics, in particular for NL query
processing.
80
Frost, R. and Peelar, S.
A New Data Structure for Processing Natural Language Database Queries.
DOI: 10.5220/0008124300800087
In Proceedings of the 15th International Conference on Web Information Systems and Technologies (WEBIST 2019), pages 80-87
ISBN: 978-989-758-386-5
Copyright
c
2019 by SCITEPRESS Science and Technology Publications, Lda. All rights reserved
Table 1: Events of type “Discover”. The full URIs of the
events, properties, and entities have been omitted here.
Event Property Entity
event1045 subject hall
event1045 object phobos
event1045 type discover ev
event1045 year 1877
event1045 location us naval observatory
event1045 implement refractor telescope 1
Table 2: Events of type “Membership”.
Event Property Entity
event1128 subject galileo
event1128 object person
event1128 type membership
2 HOW TO ACCESS OUR NLQI
Our NL interface can be accessed at the following
web site:
http://speechweb2.cs.uwindsor.ca/solarman4/
demo sparql.html
2.1 The Triplestore that is Queried
Our NLQI computes answers with respect to a triple-
store containing data about the planets, the moons
that orbit them, and the people who discovered those
moons, and when, where and with what implement
they were discovered. Note that each triple could be
equally well be represented by a row in a relational
database.
The triplestore contains triples such as the follow-
ing which represent the event #1045 in which hall
(in the role of subject) discovered phobos (in the
role of object) in 1877 (in the role of year) with the
refractor telescope 1 (in the role of implement)
at the us naval observatory (in the role of loca-
tion).
Events representing set membership are repre-
sented as follows:
The complete triplestore, which contains tens of
thousands of triples, is hosted on a remote com-
pute server using the Virtuoso software (Erling and
Mikhailov, 2010) and can be accessed by following
the link at the beginning of Section 2.
3 EXAMPLE QUERIES
Our NLQI can answer millions of queries with re-
spect to the triplestore discussed above. The NLQI
can accommodate queries containing common and
proper nouns, adjectives, conjunction and disjunction,
intransitive and transitive verbs, nested quantification,
superlatives, chained prepositional phrases containing
quantifiers, comparatives and polysemantic words. In
the following we provide an informal explanation of
how the answer is computed.
3.1 Queries Demonstrating the Range
of NL Features that our NLQI can
Accommodate
phobos spins True
phobos is a moon True
The function denoted by phobos checks to see if
e
phobos
is a member of the spin set, and secondly if
e
phobos
is a member of the moon set.
a moon spins True
every moon spins True
an atmospheric moon exists True
The function denoted by a checks to see if the in-
tersection of the set of moons and the set of spins is
non-empty. The function denoted by every checks
to see if the set of moons is a subset of the spins set.
The denotations of a and every that we use are set-
theoretic event-based versions of the denotations from
MS which uses characteristic functions.
The answer to the third query is obtained by
checking if the intersection of the atmospheric set
and the moon set is non-empty.
hall discovered True
All of the events of type “discovered” are col-
lected together and are checked to see if e
hall
is found
as the subject role value of any of them. If so, True is
returned.
when did hall discover 1877
The “year” property of the events returned by
“hall discover” (treated as “hall discovered”) are
returned.
phobos was discovered True
A New Data Structure for Processing Natural Language Database Queries
81
All of the events of type “discovered” are col-
lected together and are checked to see if e
phobos
is
found as the object role value of any of them. If so
True is returned.
earth was discovered False
Earth was not discovered by anyone, according to
our data.
did hall discover phobos True
All of the events of type “discover” are collected
together and are checked to create a pair (s, evs) for
each value of the subject attribute found in the set of
events. evs is the set of events to which the subject
attribute is related through a discovery event. Each
pair is then examined to see if the function denoted by
the object termphrase (in this case phobos) returns a
non-empty set when applied to a set (called an FDBR,
which is described later) generated from the set of evs
in the pair, and if so the subject of the pair is added
to the set which is returned as the denotation of the
verbphrase part of the query. The denotation of the
termphrase at the beginning of the query is then ap-
plied to the denotation of the verbphrase to obtain the
answer to the query.
Owing to the fact that our semantics is compo-
sitional the subject and object termphrases of the
query above can be replaced by any termphrases. For
example:
a person or a team discovered every moon
that orbits mars True
who discovered 2 moons that orbit mars
hall
“who”, “what”, “where”, “when” and “how” can
be used in place of the subject termphrase. Different
role values are returned depending on which “wh..
word is used in the query.
where discovered by galileo padua
when discovered by galileo 1610
every telescope was used to discover a
moon True (w.r.t.our data)
a moon was discovered by every telescope
False
a telescope was used by hall to discover
two moons True
which moons were discovered with two
telescopes
halimede laomedeia sao themisto
who discovered deimos with a telescope
that was used to discover
every moon that orbits mars hall
who discovered a moon with two telescopes
nicholson science team 18 science team 2
how was sao discovered blanco telescope
canada-france-hawaii telescope
how discovered in 1877 refractor telescope 1
how many telescopes were used to discover
sao 2
who discovered sao science team 18
how did science team 18 discover sao
blanco telescope canada-france-
hawaii telescope
which planet is orbited by every moon
that was discovered by
two people saturn; none
which person discovered a moon in 1877
with every telescope that was used to
discover phobos hall; none
who discovered in 1948 and 1949 with a
telescope kuiper
3.2 Queries with “Non-compositional”
Structures
We agree that natural language has non-compositional
features but believe that the non-compositionality is
mostly problematic when the objective is to give a
meaning to an NL expression (without a context). It
is less problematic when answering NL queries. As
illustrated below, the person posing the query, or the
database or triplestore can provide contexts that help
resolve much of the ambiguity resulting from non-
compositional features.
The advantages of a using a compositional seman-
tics include 1) the answer to a query is as correct
as the data from which it is derived, 2) the meaning
of sub phrases within a query can be discussed for-
mally, 3) the query language can be extended such
that all existing phrases maintain their original mean-
ings, 4) the definition of syntax and semantics in the
compositional semantics can be used as a blueprint
for the implementation of the query processor.
Some researchers have provided examples of
what they claim to be non-compositional structures
in NL. For example, Hirst (Hirst, 1992) gives the
example of the verb “depart” which he states is not
compositional because its meaning changes with the
prepositional phrase(s) which follow it, and that the
definition of compositionality needs to be modified
to include the requirement that the function used to
compose the meaning of parts must be systematic.
We claim that our semantics for verbs is systematic
as the denotations of subject and object termphrases,
WEBIST 2019 - 15th International Conference on Web Information Systems and Technologies
82
and the possibly empty list of prepositional phrases
following the verb are treated equally and are all used
in the same way to filter the set of events of the type
associated with the verb, before that set is returned as
the denotation of the verb phrase. This is illustrated
in the following queries:
who discovered bernard bond cassini
cassini imaging science team christy dollfus galileo
etc. . .
No subject, object or prepositional phrase is given
in the query, and so all events of type “discover”
are returned by the verbphrase and the denotation of
the word who picks out the subjects from those events.
where discovered io padua
No subject, or prepositional phrase is given in the
query, and so all events of type “discover” are con-
sidered and filtered by the denotation of the object
termphrase io and then, those that pass the filter are
returned by the verbphrase and the word where picks
out the location from those events.
who discovered in 1610 galileo
No subject or object is in the query so all events of
type “discover” are considered and only those with at-
tribute “date” equal to 1610 pass the filter and then the
denotation of the word who selects the subject which
is returned.
In our semantics, the subject and object
termphrases are treated as filters as are all preposi-
tional phrases, as shown in the following example:
who discovered every moon that orbits
mars with one telescope or a
moon that orbits Jupiter with a telescope
one. ; none. ; none. ; bernard galileo kowal
melotte nicholson perrine science
team 1 sci-
ence team 2 ; hall ; hall ; none.
Several results are returned because the query is
syntactically ambiguous.
where discovered in 1610 padua
how discovered in padua galilean telescope 1
3.3 Extensions to the Semantics
Some phrases containing nested quantifiers are
given by some researchers, as examples of non-
compositionality. For example: “a US diplomat was
sent to every capital” is often read as having two
meanings which can only be disambiguated by addi-
tional knowledge. We argue that the person posing a
query can express the query unambiguously if they
are familiar with quantifier scoping conventions used
by our processor, as illustrated in the following:
christy or science team 19 or
science team 20 or science team 21
discovered every moon that orbits pluto
False
In our semantics, quantifier scoping is always
leftmost/outermost, and an unambiguous query can
be formulated as follows:
every moon that orbits pluto was
discovered by christy or
science team 19 or science team 20 or
science team 21 True
Some examples of non-compositionality involve
polysemantic superlative words such as “most” in, for
example:
“Who discovered most moons that orbit P. Where P
is a planet.
If “most” is treated as “more than half” then:
who discovered most moons that orbit mars
hall
Because our semantics currently allows only this
reading. However, the answer to the alternate reading
“who discovered the most moons that orbit P i.e.
more than anyone else who discovered a moon that
orbits P. Could be obtained in our semantics by
comparing all of the (ent, evs) pairs returned by the
verbphrase to see which subject is paired with most
objects. We are currently working on this and other
extensions to our semantics.
how was every moon that orbits saturn
discovered cassini
reflector telescope 1 aerial telescope 1 refrac-
tor telescope 4 etc. . .
It may be surprising that cassini is returned in the
answer since it is not a telescope, but is instead a
spacecraft. However, since it was used to discover
at least one moon that orbits saturn, it is considered
to have fulfilled the implement role and is encoded as
such in the triplestore.
A New Data Structure for Processing Natural Language Database Queries
83
4 THE FDBR: A NOVEL DATA
STRUCTURE FOR NATURAL
LANGUAGE QUERIES
4.1 Montague Semantics
All quantifiers, such as “a”, “every” and “more than
two” are treated in MS as functions which take two
characteristic functions of sets as arguments and re-
turn a Boolean value as result. Our modifications
to MS are to use sets of entities instead of predi-
cates/characteristic functions of those sets, and to pair
sets of events with each entity; the set of events paired
with an entity justify the entity’s inclusion in the de-
notation. For example:
kpropernounk =
λp.{(e, evs) | (e, evs) p & e =
the entity associated with the proper noun}
kspinsk = {(e
phobos
, {ev
1360
}), (e
deimos
, {ev
1332
}),
etc. . .}
Therefore,
kphobos spinsk = kphobosk kspinsk
= λs. {(e, evs)|(e, evs) s & e = e
phobos
}
{(e
phobos
, ev
1360
), (e
deimos
, ev
1332
), . . .}
= {(e
phobos
, ev
1360
)}
kak = λmλs. {(e
1
, evs
2
) | (e
1
, evs
1
)) m
& (e
2
, evs
2
) s & e
1
= e
2
})
ka moon spinsk = {(e
phobos
, ev
1360
),
(e
deimos
, ev
1332
), etc. . .}
Note that the events evs paired with the entities
returned in the denotation of “was every moon that
orbits Saturn discovered” are the events representing
membership of those entities of type moon in the ob-
ject value of events of type discovery. This enables
additional data to be accessed from those events, as
illustrated in the last example query in the previous
section.
4.2 The FDBR
In order to generate the answer to “hall discovered
every moon that orbits mars”, keveryk is applied to
kmoon that orbits marsk (i.e. the set of moons that
orbit mars), as first argument, and the set of entities
that were discovered by hall, as the second argu-
ment. Our semantics generates this set from the set
of events of type “discover” where the subject role is
the entity associated with hall, as discussed below:
Every set of n-ary events (i.e. events with n roles)
of a given type, e.g. discovery, defines n
2
n binary
relations. For example, for discovery events:
discover rel
subjectobject
discover rel
subjectyear
discover rel
subjectimplement
. . .
discover rel
objectsubject
discover rel
objectyear
discover rel
objectimplement
. . .
discover rel
yearsubject
discover rel
yearobject
discover rel
yearimplement
. . .
etc. . . to 20 binary relations for the set of discovery
events or an 5-ary discovery relation. For example:
discover rel
subjectobject
=
{(ev
1045
, e
hall
, e
phobos
), (ev
1046
, e
hall
, e
deimos
), etc . . .}
If we collect all of the values from the range of a re-
lation that are mapped to by each value v from the
domain (i.e. the image of v under the relation r) and
create the set of all pairs (v, image o f v), we obtain a
function defined by the relation r, i.e. the FDBR. For
example:
FDBR(discover rel
subjectobject
)
= {(e
hall
, {(e
phobos
, {ev
1045
}), (e
deimos
, {ev
1046
})}),
etc. . .}
It is these functions that are created, and
used, by the denotation of the transitive
verb associated with the type of the events.
For example in calculating the value of
kwho discovered every moon that orbits marsk,
keveryk is applied to the set of entities which is
the denotation of “moon that orbits mars” (i.e
{(e
phobos
, {ev
1045
}), (e
deimos
, {ev
1046
})} ) and all of
the images that are in the second field of the pairs
in FDBR(discover rel
subjectobject
). For the pair
(e
hall
, {(e
phobos
, {ev
1045
}), (e
deimos
, {ev
1046
})}),
keveryk returns the non-empty set
{(e
phobos
, {ev
1045
}), (e
deimos
, {ev
1046
})}, and the
value in the first field, i.e. e
hall
, is subsequently
returned with the answer to the query.
The various FDBRs are used to answer different
types of queries. For example:
who discovered phobos and deimos hall
uses FDBR(discover
rel
subjectobject
)
where discovered by galileo padua
uses FDBR(discover rel
locationsubject
)
how discovered in 1610 or 1855
galilean telescope 1
uses FDBR(discover rel
implementyear
)
WEBIST 2019 - 15th International Conference on Web Information Systems and Technologies
84
5 HANDLING PREPOSITIONAL
PHRASES
Prepositional phrases (PPs) such as “with a telescope”
are treated similarly to the method above, except that
the termphrase following the preposition is applied to
the set of entities that are extracted from the set of
events in the FDBR function, according to the role
associated with the preposition. The result is a “fil-
tered” FDBR which is further filtered by subsequent
PPs.
6 QUANTIFIERS AND EVENTS
In 2015, Champollion (Champollion, 2015) stated
that, at that time, it was generally thought by lin-
guists that integration of Montagovian-style composi-
tional semantics and Davidsonian–style event seman-
tics (Parsons, 1990; Davidson, 1967) was problem-
atic, particularly with respect to quantifiers. Cham-
pollion did not agree with that analysis and presented
an integration which he called “quantificational event
semantics” which he claimed solved the difficulties
of integration by assuming that verbs and their pro-
jections denote existential quantifiers over events and
that these quantifiers always take lowest possible
scope.
In this paper, we borrow much from Mon-
tague Semantics (MS), Davidsonian Event Semantics,
and Champollion’s Quantificational Event Semantics.
However, we provide definitions of our denotations
in the notation of set theory, which improves compu-
tational efficiency and, we believe, simplifies under-
standing of our denotations. We also believe that our
semantics is intuitive, systematic, and compositional.
7 OUR APPROACH WITH
RELATIONAL DATABASES
Our NLQI could be easily adapted for use with con-
ventional relational databases. Each row in a relation
Rel can be thought of as representing an event of type
Rel, and each column name can be thought of as a
role name. The event itself would serve as the pri-
mary key, and only the triple retrieval function would
need to be modified. This architecture allows the de-
notations to remain unchanged and yet still work with
different types of databases.
8 IMPLEMENTATION OF OUR
NLQI
We built our query processor as an executable at-
tribute grammar using the X-SAIGA Haskell parser-
combinator library package (Frost et al., 2008). The
collect function which converts a binary relation to
an FDBR is one of the most compute intensive parts
of our implementation of the semantics. However, in
Haskell, once a value is computed, it can be made
available for future use. We have developed an al-
gorithm to compute FDBR(rel) in O(n log n) time,
where n is the number of pairs in rel. Alternatively,
the FDBR functions can be computed and stored in
a cache when the NLQI is offline. Our implemen-
tation is amenable to running on low power devices,
enabling it for use with the Internet of Things. A ver-
sion of our query processor exists that can run on a
common consumer network router as a proof of con-
cept for this application. The use of Haskell for the
implementation of our NLQI has many advantages,
including:
1. Haskell’s “lazy” evaluation strategy only com-
putes values when they are required, enabling
parser combinator libraries to be built that can
handle highly ambiguous left-recursive grammars
in polynomial time.
2. The higher-order functional capability of Haskell
allows the direct definition of higher-order func-
tions that are the denotations of some English
words and phrases.
3. The ability to partially apply functions of n argu-
ments to 1 to n arguments allows the definition
and manipulation of denotation of phrases such as
“every moon”, and “discover phobos”.
4. The availability of the hsparql (Wheeler, 2009)
Haskell package enables a simple interface be-
tween our semantic processor and SPARQL end-
points to our triplestores.
9 RELATED WORK
Orakel (Cimiano et al., 2007) is a portable NLQI
which uses a Montague-like grammar and a lambda
calculus semantics. Our approach is similar in this re-
spect. Queries are translated to an expression of first
order logic enriched with predicates for query and nu-
merical operators. These expressions are translated to
SPARQL or F-Logic. Orakel supports negation, lim-
ited quantification, and simple prepositional phrases.
YAGO2 (Hoffart et al., 2013) is a semantic knowl-
edge base containing reified triples extracted from
A New Data Structure for Processing Natural Language Database Queries
85
Wikipedia, WordNet and GeoNames, representing
nearly 0.5 billion facts. Reification is achieved by tag-
ging each triple with an identifier. However, this is
hidden from the user who views the knowledge base
as a set of “SPOTL quintuples, where T is for time
and L for location. The SPOTLX query language is
used to access YAGO2. SPOTLX can handle queries
with prepositional aspects involving time and loca-
tion. However, no mention is made of chained com-
plex PPs.
Alexandria (Wendt et al., 2012) is an event-based
triplestore, with 160 million triples (representing 13
million n-ary relationships), derived from FreeBase.
Alexandria uses a neo-Davidsonian (Parsons, 1990)
event-based semantics. In Alexandria, queries are
parsed to a syntactic dependency graph, mapped to
a semantic description, and translated to SPARQL
queries containing named graphs. Queries with sim-
ple PPs are accommodated. However, no mention is
made of negation, nested quantification, or chained
complex PPs.
The systems referred to above have made sub-
stantial progress in handling ambiguity and match-
ing NL query words to URIs. However, they ap-
pear to have hit a roadblock with respect to natural-
language coverage. Most can handle simple PPs such
as in “who was born in 1918” but none can handle
chained complex PPs, containing quantifiers, such as
“in us naval observatory in 1877 or 1860”.
Blackburn and Bos (Blackburn and Bos, 2005)
implemented lambda calculus with respect to natural
language, in Prolog, and (Van Eijck and Unger, 2010)
have extensively discussed such implementation in
Haskell. Implementation of the lambda calculus for
open-domain question answering has been investi-
gated by (Ahn et al., 2005). The SQUALL query lan-
guage (Ferre, 2012; Ferr
´
e, 2013) is a controlled natu-
ral language (CNL) for querying and updating triple-
stores represented as RDF graphs. SQUALL can re-
turn answers directly from remote triplestores, as we
do, using simple SPARQL-endpoint triple retrieval
commands. It can also be translated to SPARQL
queries which can be processed by SPARQL end-
points for faster computation of answers. SQUALL
can handle quantification, aggregation, some forms of
negation, and chained complex prepositional phrases.
It is also written in a functional language. However,
some queries in SQUALL require the use of variables
and low-level relational algebraic operators (see for
example, the queries on page 118 of (Ferr
´
e, 2013)).
10 CONCLUDING COMMENTS
We are confident that, after we accommodate nega-
tion, our compositional semantics is appropriate for
most queries that are likely to be asked of data stores
containing everyday knowledge. The FDBR datas-
tructure presented in this paper can be used to handle
many kinds of complex language features, including
chained prepositional phrases and superlatives. The
way quantification is handled within the semantics is
consistent with other work in this area, as discussed
in Section 6. The approach chosen is flexible enough
that it can accommodate queries to both relational and
non-relational types of databases, including Semantic
Web triplestores. It is also suitable for use on low
power devices, which may be useful for applications
on the Internet of Things (IoT).
In the future, we plan to scale up the capability of
our NLQI further to access massive data stores such
as DBpedia. To achieve this goal, we plan to accel-
erate the FDBR generation process using specialized
acceleration hardware, such as FPGAs and GPUs.
REFERENCES
Ahn, K., Bos, J., Kor, D., Nissim, M., Webber, B. L., and
Curran, J. R. (2005). Question answering with qed at
trec 2005. In TREC.
Blackburn, P. and Bos, J. (2005). Representation and infer-
ence for natural language. A first course in computa-
tional semantics. CSLI.
Champollion, L. (2015). The interaction of compositional
semantics and event semantics. Linguistics and Phi-
losophy, 38(1):31–66.
Cimiano, P., Haase, P., Heizmann, J., and Mantel, M.
(2007). Orakel: A portable natural language interface
to knowledge bases. Technical report, Technical re-
port, Institute AIFB, University of Karlsruhe.
Davidson, D. (1967). The logical form of action sentences.
Erling, O. and Mikhailov, I. (2010). Virtuoso: Rdf support
in a native rdbms. In Semantic Web Information Man-
agement, pages 501–519. Springer.
Ferre, S. (2012). Squall: A controlled natural language for
querying and updating rdf graphs. In proc. of CNL
2012, pages 11–25. LNCS 7427.
Ferr
´
e, S. (2013). Squall: a controlled natural language as
expressive as sparql 1.1. In International Conference
on Application of Natural Language to Information
Systems, pages 114–125. Springer.
Frost, R. and Launchbury, J. (1989). Constructing natu-
ral language interpreters in a lazy functional language.
The Computer Journal, 32(2):108–121.
Frost, R. A., Hafiz, R., and Callaghan, P. (2008). Parser
combinators for ambiguous left-recursive grammars.
In International Symposium on Practical Aspects
WEBIST 2019 - 15th International Conference on Web Information Systems and Technologies
86
of Declarative Languages, pages 167–181. Springer
LNCS Volume 4902.
Hirst, G. (1992). Semantic interpretation and the resolution
of ambiguity. Cambridge University Press.
Hoffart, J., Suchanek, F. M., Berberich, K., and Weikum,
G. (2013). Yago2: A spatially and temporally en-
hanced knowledge base from wikipedia. Artificial in-
tell., 194:28–61.
Parsons, T. (1990). Events in the Semantics of English, vol-
ume 5. Cambridge, Ma: MIT Press.
Van Eijck, J. and Unger, C. (2010). Computational seman-
tics with functional programming. Cambridge Univer-
sity Press.
Wendt, M., Gerlach, M., and D
¨
uwiger, H. (2012). Lin-
guistic modeling of linked open data for question an-
swering. proc. of interact. with Linked Data (ILD
2012)[37], pages 75–86.
Wheeler, J. (2009). The hsparql pack-
age. In The Haskell Hackage Repository.
http://hackage.haskell.org/package/hsparql-0.1.2.
A New Data Structure for Processing Natural Language Database Queries
87