Natural Semantic Metalanguage (NSM)

Semantic primes are the vocabulary of the Natural Semantic Metalanguage. NSM grammar specifies how primes can be combined in ways that make sense and appear to be possible in all languages. This table displays the English exponents of the primes and some of their basic combinatorial possibilities. Many other, more complex combinations are possible, especially using operators and connectives like NOT, CAN, MAYBE, IF, and BECAUSE, and drawing on the complement-taking properties of KNOW, WANT and THINK. Primes can have two or more exponents (allolexes) in a given language, e.g. other and else in English. In some languages certain combinations of primes are expressed by a portmanteau, e.g. in Polish the combination ‘like this’ is expressed by a single word tak.

Using NSM in explications

  • Whenever possible, compose explications entirely in semantic primes. Don’t include any complex English-specific words, no matter how common and seemingly basic; e.g. andormake.
  • Ensure that you use semantic primes only in allowable syntactic frames. Don’t resort to non-NSM syntax even if English grammar allows it. For example, avoid ‘do something about it’ and ‘feel good/bad about something’; semantic primes DO and FEEL do not have any “about”-valency. Likewise, avoid ‘for a moment’; semantic prime MOMENT is incompatible with duration.
  • Avoid the following non-universal constructions: relative clauses (e.g. someone   who   I   know   well), comparative (e.g. better   thanmore   than),   indirect speech (e.g. this   someone   said   that   …).
  • Words of a similar semantic type tend to follow a common semantic template. Some explications require use of semantic molecules, in addition to primes. Semantic molecules are a well- defined set of non-primitive meanings that function as units in the meanings of more complex concepts.

Cliff Goddard <>4 January 2011

Relational SubstantivesKIND, PART
EvaluatorsGOOD, BAD
DescriptorsBIG, SMALL
ActionsEventsMovementDO, HAPPEN, MOVE
Life and DeathLIVE, DIE
Logical ConceptsNOT, MAYBE, CAN, BECAUSE, IF
Intensifier, AugmenterVERY, MORE



I don’t know, I want you to do/know something, something bad can happen to me/you, someone like me


something is happening here now


something is happening here now


this someone, the same someone, someone else, this other someone


(in) this place, (in) the same place, somewhere else, (in) this other place, in some places, in many places,

in the place where …


(at) this time, (at) the same time, at another time, at this other time, at some times, at many times,

at the time when …


this thing, the same thing, something else, this other something


this part, the same part, another part, this other part, part of something, part of someone’s body, part of a place, this thing has two/many parts


this kind, the same kind, another kind, this other kind, something/someone of one/two/many kinds, people of one/two/many kinds


many people, some people, people think like this, people can say


part of someone’s body, two kinds of bodies, body of one kind


many words, other words, one word, say something with (not with) words, say these words, these words say something


this someone (something), these people, at this time, in this place, this kind, this part, because of this, it is like this


the same someone, the same thing, the same part, the same kind, at the same time, in the same place, someone says/does/thinks/knows/wants/feels the same


someone else, something else, at another time, somewhere else, other parts, other kinds, this other part, this other kind, this other someone, this other thing


all people, all things, all parts, all kinds, at all times, in all places


one someone, one thing, one part, one kind, one of these things/people, something of one kind, at one time, in one place, one more thing


many people, many things, many parts, many kinds, at many times, in many places much something of this kind (e.g. water), much/many more


some people, some things, some parts, some kinds, at some times, in some places, some of these things/people


two things, two parts, two kinds,

two of these things/people, two more things


few people, few things

a little something of this kind (e.g. water)


something good, someone good, good people, something good happens, do something good (for someone), feel something good, this is good, it is good if ...


something bad, bad people, something bad happens, do something bad (to someone), feel something bad, this is bad, it is bad if ...


this is true, this is not true


something big, a big place, a big part


something small, a small place, a small part


very big, very small, very good, very bad, very far, very near, a very short time, a very long time


something happens

something happens to someone something happens to something something happens in a place

someone does something (to someone else) someone does something to something (with

something else)

someone does something with someone else

someone does something good for someone else

someone says something (to someone) someone says something (good/bad)

(about someone/something) someone says something like this: “- -”

someone says something (not) with words


someone wants something,

someone wants to do/know/say something, someone wants someone else to do/know


someone wants something to happen

someone feels something (good/bad) (in part of the body)

someone feels like this

someone feels something good/bad towards someone else

someone thinks (something good/bad) about someone/something

someone thinks like this: “- -”, many people think like this: “- -”

(at this time) someone thinks that ...


someone sees someone/something (in a place)

someone hears something

someone knows something (about someone/something)

someone knows when/where/who ... someone knows that ...

people can know this

someone knows someone else (well)

BE (specificational)

this someone is someone like me this is something of one kin

this is something big/small

someone can say who this someone is someone can say what kind of thing

this is …


there is something in this place there is someone in this place there are two/many kinds of …


someone moves (in this place) something moves in this place parts of this someone’s body move

BE (locational)


someone has something (many things) someone has something of this kind

someone is in a place something is in a place someone is with someone else

something touches something else (in a place)

something touches someone (part of this someone’s body)

someone touches someone else (part of this other someone’s body)


someone lives for a long time someone lives in this place many people live in this place

someone lives with someone else


someone dies at this time all people die at some time


someone wants more, someone does more, someone wants to know/say more about it, one more, two more, many more not living anymore, not like this anymore


I don’t know, I don’t want this, someone can’t do this, it is not like this, not good, not bad, not because of anything else


someone can (can’t) do something someone can’t not do something something (good/bad) can happen it can be like this


because of this/it, it happened because this someone did something before


do it like this, move like this, happen like this, do it in this way

think like this: “- -”, it is like this: …, like/as this someone wants someone like me


maybe it is like this, maybe it is not like this, maybe someone else can do it


if it happens like this for some time, ..., if you do this, ..., if someone does something like this, ...


before this, some time before, a short time before, a long time before


for a short time, a short time before, a short time after


it happens for some time, someone does this for some time


after this, some time after, a short time after, a long time after


for a long time, a long time before, a long time after


it happens in one moment


above this place


near this place, near someone


inside this something, inside part of someone’s body


below this place


far from this place


on this side, on the same side, on one side, on two sides, on all sides


Key References

  • Wierzbicka, A. 1996. Semantics: Primes and Universals. OUP. ¦ Goddard, C. & Wierzbicka, A. Eds. 2002. Meaning and Universal Grammar. Vols I and II. Benjamins. ¦ Peeters, B. Ed. 2006. Semantic Primes and Universal Grammar: Evidence from the Romance Languages. Benjamins. ¦ Goddard, C. Ed. 2008. Cross-Linguistic Semantics. Benjamins. ¦ Goddard, C. 2011. Semantic Analysis [2nd edn]. OUP. ¦ Goddard, C. & Wierzbicka, A. In press. Words & Meanings. OUP. ¦ NSM Homepage:

Using NSM in explications

  • Whenever possible, compose explications entirely in semantic primes. Don’t include any complex English-specific words, no matter how common and seemingly basic; e.g. and, or, make.
  • Ensure that you use semantic primes only in allowable syntactic frames. Don’t resort to non-NSM syntax even if English grammar allows it. For example, avoid ‘do something about it’ and ‘feel good/bad about something’; semantic primes DO and FEEL do not have any “about”-valency. Likewise, avoid ‘for a moment’; semantic prime MOMENT is incompatible with duration.
  • Avoid the following non-universal constructions: relative clauses (e.g. someone   who   I   know   well), comparative (e.g. better   than, more   than),   indirect speech (e.g. this   someone   said   that   …).
  • Words of a similar semantic type tend to follow a common semantic template. Some explications require use of semantic molecules, in addition to primes. Semantic molecules are a well- defined set of non-primitive meanings that function as units in the meanings of more complex concepts.

Cliff Goddard <>4 January 2011


Are there Semantic Primes in Formal Languages?

Johannes Fahndrich, Sebastian Ahrndt and Sahin Albayrak¨

DAI Lab, Berlin Institute of Technology, Berlin, Germany


Abstract. This paper surveys languages used to enrich contextual information with semantic descriptions. Such descriptions can be e.g. applied to enable reasoning when collecting vast amounts of row data in domains like smart environments. In particular, we focus on the elements of the languages that make up their semantic. To do so, we compare the expressiveness of the well-known languages OWL, PDDL and MOF with a theory from linguistic called the Natural Semantic Metalanguage.

Keywords: Context Description Languages, Contextual Reasoning, OWL, PDDL, NSM

  1. Introduction

Intelligent environments are made up of multiple pervasive or ubiquitous devices that provide a service to the user. One key indicator of such environments is the ability to adapt to changes. The changes are implied by external or internal influences like the introduction or removal of devices or changing application goals [20]. We expect that intelligent environments react to such changes and adapt themselves in a way that the service provided are still available for the users. More than a decade ago R.J. Sternberg [22] still emphasises this specifying that intelligence is the ability to adapt to changes in environments (to distinguish between the environment itself we refer to this as context). This point of view implies that an environment becomes more intelligent if it can cope with more or bigger changes in the context. To be able to adapt to contextual changes a cognition is needed to be aware of the actual context and appearing changes. We focus our analysis to environments where such cognition is available. That means, that there exist at least one entity able to perceive the context and able to communicate the actual perception to other entities in the environment.

One can distinguish two types of contextual information: The defined context and the derived context [13]. In both cases, the devices making up the intelligent environment have to agree on a language to interpret the data collected by the sensing devices. In a defined context (e.g., a specific application) this language can be given to the environment by a domain model. Another approach is to use semantic languages to annotate contextual information and use reasoner that derive knowledge or facts from this annotations. This approach is called derived context. Here every device has its own local model of the environment, without having to agree on a global context model providing information about all devices. Derived context is created by finding patterns in raw data form the sensing devices of an intelligent environment and annotate them with the given semantic language. A reasoner then reasons upon this annotated information to transform the information into a local domain model of the device. This emphasises the requirement to agree upon a semantic language used for the annotation. Furthermore, it underlines why no model of the whole context is needed.

Languages to describe semantic have been subject to research in many research areas. Bikakis et al. [1] surveys semantic based approaches and applicable reasoning methods in the domain of ambient intelligence. Two of them are the Semantic Web community and the Agent community. Both have developed a quasi standard language to describe semantics. In the semantic web community the Web Ontology Language[1] [17] (OWL) is been widely used. The agent community uses Planning Domain Definition Language[2] [7] (PDDL) to describe their planning problems. This paper will examine the fundamental concepts making up those two languages. Additionally the study includes the Meta Object Facility [6] (MOF) as a meta-language for artificial languages. We compare these approaches with a theory form linguistics named the Natural Semantic Metalanguage [8] (NSM). NSM states that every naturally developed language is based on 63 semantic concepts.

The paper is structured in the following way: In Section 2 introduces NSM and the basic concept of semantic primes in a nutshell. Furthermore, it describes the three semantic description languages and the difference between the languages. Section 3 takes such insights into account and compares the languages in a more detailed way. Afterwards, Section 4 wraps-up the paper with a discussion of the results.

  1. Semantic Primes

The Natural Semantic Metalanguage (NSM) is a linguistic theory originated in the early 1970s [25]. It states that each meaning of an concepts created in a natural language can be represented using a set of atomic terms—so-called universal semantic primes. These primes have an indefinable word-meaning and can be identified in all natural languages [9]. In conjunction with associated grammatical properties NSM presents a decompositional system able to describe all concepts build in the appropriate language. Here, an expression is decomposed into less complex concepts, where the process ends if the expression is decomposed to the atomic level of semantic primes which can not be analyzed further. One can imagine that the decomposition builds a tree, where all leafs are semantic primes [27]. Consequently for each natural language a metalanguage exist that consist of the semantic primes in the specific syntax and their appropriated grammatical properties. About 63 semantic primes exist that can be divided into 16 categories [26].

As well as natural languages, formal defined artificial languages are based on a meta-language like the Meta Object Facility. This leads to the implication that the concepts defined in artificial languages are semantic primes and that such primes can be compared among different languages. Since the bag of semantic primes presented

 by NSM is empirically well-researched, this work tries to compare three artificial languages utilising this bag of primes. For this comparison, we take the purpose and concepts of the languages into account and match the available primes with each other as foundation to discuss potentially missing primes in the languages.

Table 1. List of semantic primes with no equivalent found in the other languages.

CategorySemantic prime
QuantifiersTWO, MUCH/MANY
EvaluatorsGOOD, BAD
DescriptorsBIG, SMALL
Mental predicatesTHINK, KNOW, WANT, FEEL, SEE, HEAR

Actions, events, movement, contact





Intensifier, augmenterVERY, MORE
Logical conceptsMAYBE

Table 1 lists all semantic primes presented by NSM not matchable with any concept in the examined artificial languages. In the following we will introduce the artificial languages and the list for each language, which semantic primes are used.

The Web Ontology Language (OWL) is a semantic markup language to create structured knowledge representations and enrich them with semantics. OWL is a W3C standard since 2004 and has been continuously developed since [11]. OWL is an extension of the Resource Description Framework [15] and has become one of the most used language to describe knowledge for AI. Since OWL is meant to describe structured knowledge the concepts used are abstract. Table 2 list all equivalents found in comparison with NSM primes.

Table 2. List of semantic primes with and equivalent found in OWL.

Category        Semantic primeOWL
Substantive Iself.entry
Relational      KINDowl:SubClassOf
substantives PARTowl:topObjectProperty
Determiners THISowl:entityURI
THE SAMEowl:equivalentClass

The Planning Domain Definition Language (PDDL) is a first-order logic based language defined as an extended BNF [7]. Commonly, it is used to provide a standardized way to describe planning problems and the associated domains. The syntax allows to define among others actions, effects, quantifications and constraints and was intended to enable developers to describe the ”physics” of a domain. Given such a description the reasoner uses a goal defined in PDDL to search for a plan that satisfies all constraints, requirements and preconditions. The concepts which are equivalent to semantic primes are listed in Table 3.

Table 3. List of semantic primes with and equivalent found in PDDL.

CategorySemantic primePDDL
SubstantiveSOMETHING/ THING:define
ExistenceTHERE IS:exists
 A LONG TIME:maintain
 A SHORT TIME:wait-for
Logical concepts NOT:not

The Meta Object Facility (MOF) has been introduced by the Object Management Group and is formally defined e.g. by Smith et al. [21]. MOF has been developed to model the structure and behaviour of entities in software development. For example, UML[3] implements MOF. Since MOF is quite abstract, the meta language like OWL mostly has structural semantic primes. Table 4 list all equivalents.

Table 4. List of semantic primes with and equivalent found in MOF.

CategorySemantic primeMOF
RelationalKINDtype, extent
DeterminersTHE SAMEelement.equals
LocationBE (SOMEWHERE)link
ExistenceTHERE ISelement
Life and deathLIVEcreate
Logical concepts CANoperation
  1. Comparison of Primes

The compared languages introduce additional concepts that are domain specific and which are not part of the semantic primes (e.g., ‘OWL:VERSIONINFO’). Depending on the purpose of the language those additional concepts change. OWL for example was created to describe shared conceptualizations where versioning and backward compatibility is an issue. But from the theory of NSM those concepts could be described using the other semantic primes. Thus they are merely shortcuts. There are multiple extensions to those languages for special cases like the Semantic Web Rule Language (SWERL) [14], which introduces rules to OWL. We now discuss the 16 categories of semantic primes to analyse why such concepts do or do not exist.

Substantives are the first category. In natural language these semantic primes are used to distinguish actors and to separate humans from other things. To describe meaning, humans often reduce description of properties of things to the relation to humans or more precise them self [27]. For example, to describe the concept ‘mouse’ a semantic, context independent description most likely rely on a degree in Biology. Describing a mouse so that a na¨ıve reader of the description might understand it, the description can refer to the potential context of the reader. In natural languages these readers are other humans, which implies that most description in NSM are in the context of humans and their relation to things. The distinction of ‘YOU’ and ‘I’ is needed if roles are described, i.e. in negotiation or contracting. Non of the reviewed language has a concept for ‘PEOPLE’. On the one hand semantic description in these languages are thought for artificial reasoners and there a concept of ‘PEOPLE’ is not needed to describe most concepts. On the other hand, in the area of HCI the distinction of artificial agents and human agents can be of some concern and with that the concept of ‘PEOPLE’ might be required.

The category of relational substantives are well represented in two of the languages, except PDDL as it does not use type hierarchies to define domains. That means that PDDL does not semantically aggregate all instances of one ‘KIND’. In semantic descriptions ‘KIND’ is for example used to describe water: ’something of one kind’ [10].

Quantifiers are represented in all tree languages. The exception is the fuzzy representation of ‘MUCH/MANY’ and ‘SOME’. There is a need to enable fuzziness in semantic description languages as motivated by Stoilos et al. [23].

The category of Evaluators is not represented in any language. PDDL from version 2.1 features numeric fluent to describe e.g., cost for actions, which could be interpreted depending the metric. Here an implicit metric is given to the reasoner, e.g. the plan minimization metric [7]. This metric states that the value to be optimized is better the smaller the value. Stating that less is better. We argue that such a metric can be explicitly formalized in the description itself and to define what is ‘BAD’ or ‘GOOD’ those concepts need to be part of the description language.

Descriptors are not represented in all three analyzed language. ‘BIG’ and ‘SMALL’ are fuzzy values and are defined in an description. For example, Wierzbicka [27] describes mice as small in the following way: ’They are very small. A person can hold one easily in one hand’. Giving a example on what small means and a relation to something every reader of the explanation knows: the size of a hand. In relation to some reference or as constant like the semantic primes ‘A LONG TIME’ and ‘A SHORT TIME’ these can be used to describe relations in e.g. size, intensity, power or time. We can imagine that for example a timeout can be explained by defining ‘A LONG TIME’ as the maximum timeout. We argue that if the semantic description is used by a reasoner to create a heuristic, a metric needs to be defined and with that semantic descriptors are needed which classify the value which is subject to the metric. Further we argue that ‘LONG’ and ‘SHORT’ should be part of the descriptors since they are descriptors which could be used in addition to time with other concepts like spacial distances.

Mental predicates are not represented in all three languages doe to the fact that these predicates are based on the senses of human being. We separate mental predicates in two groups: The first group is based on human cognition, ‘FEEL’, ‘SEE’ and ‘HEAR’. First of all there are two of the senses missing: ‘SMELL’ and ‘TASTE’. Additionally for the domain of intelligent environments and if sensors needs to be described, such predicates can be used. Of cause this is specific for humans, since not every agent involved in an intelligent environment has such cognitive functions. Additionally there could be sensors which extend the human cognitional like a barometer, altimeter or a localization like GPS, which should then be incorporated in the semantic primes as well. Even thou these concept are not used in the analysed modelling languages, they are part of the semantic description of many fundamental concepts like ’HANDS’ [27]. The second group of mental predicates is the mental state of mind: ‘THINK’, ‘KNOW’, ‘WANT’. These are philosophical terms and rarely used in artificial languages. Braubach et al. [2] e.g. describe a Believe, Desire, Intent (BDI) paradigm for agents. Here ‘I believe’ is considered a subset of ‘I think’ [26]. In the BDI paradigm ‘believe’ can be mapped to the semantic prime ‘KNOW’, since it represents the knowledge of the agent and ‘desire’ can be mapped to ‘WANT’ since it describes the internal goals of the agent. But in our analysed languages, all of these concepts are missing.

Speech is - at first - the category which hold one of the basic logical operators ‘TRUE’. All three languages use an implicit representation of the concept ‘TRUE’ since they assume that a reasoner interprets an axiom as fundamentally true. PDDL e.g. ‘define a model to be an interpretation of a domain’s language that makes all its axioms true’ [16]. Thus here again it can be argued to explicitly describe such truth values and with that add a semantic prime to the metalanguage. But we think that ‘TRUE’ should be caped in the category ‘Logical concepts’. Further we use ‘WORDS’ as basic building blocks for our description, and thus need a semantic prime for them. ‘SAY’ has been represented in a formal way as agent communication speech acts [4] and could be directly part of the metalanguage.

Semantic primes in the category Action are often defined in an context dependent manner, where the semantic is given by the reasoner of the evaluating the axioms. In PDDL for example the blocks world defines‘MOVE(A,B,T)’ [12]. NSM proposes to add such primes to the metalanguage, to be able to describe events, movement and actions.

The semantic primes in the category possession e.g., ‘HAVE’ can be seen as the specialization of composition and aggregation of the semantic prime ‘PART’. The specification ‘BE’ denotes a location where something is located and at the same time to be of one type.

Life and death is a category which is not subject to research in formal languages because computer systems rarely need a concept of death or living. Semantically there are many things that can not be described without the concept of ‘LIVE’ and ‘DEATH’. In agent communication e.g.m agents send ‘alive messages’ to other agents, where the interpretation is left to the programmer of the agent.

Time has found its way into almost every formal language. Even a own logic—the Temporal logic—which is a kind of modal logic has been created to model something like: ‘I am hungry until I eat something’. In formal languages time has often been included into the language e.g., PDDL from version 2.1.

The category space is subject to research and is formulated in contextual models like the CORBA-ONT [3]. Nevertheless non of the surveyed languages presents primitive elements to describe special properties. The fact that a OWL ontology is required, shows that such semantic primes are necessary for the modelling of contexts. The same can be argued for the semantic primes ‘BE (SOMEWHERE)’ of the category location. Other fuzzy primes like ‘NEAR’ or ‘FAR’ are again hard to grasp in a formal language.

Logical concepts are of cause part of most formal languages. The hurdle primes are again the fuzzy ones: ‘MAYBE’ and ‘CAN’. To describe the meaning of probability, those primes could be part of a language like they are in epistemic logic [24].

Intensifier can be modeled as lexical functions [18] and fuzzy decision systems have been subject to research [19] and thus to make their semantic explicit intensifier should be part of the metalanguage.

Similarities are a huge research area and measures have been studied in depth [5]. The here developed methods like recommender systems try to find entities which are alike. Those methods try do define the prime ’SAME’ for different domains.

  1. Conclusion

We have analysed three common semantic description languages and compared their meta languages with the set of semantic primes taken from NSM. We have found that already many of the semantic primes are part of the three formal description languages depending on their focus. The semantic primes that are not yet part of the description languages has been collected in Table 1. Future work will include an in-depth analysis of those primes. Here we want to examine which primes are useful for formal languages and define a set-theoretic semantic for each of them.

Acknowledgement: Research in this paper has been financed by the Schaufenster

Elektormobilitat in the project 16SBB007A.¨


  1. Antonis Bikakis, Theodore Patkos, Grigoris Antoniou, and Dimitris Plexousakis. A survey of semantics-based approaches for context reasoning in ambient intelligence. In Constructing Ambient Intelligence, pages 14–23. Springer, 2008.
  2. Lars Braubach, Alexander Pokahr, and Daniel Moldt. Goal representation for bdi agent systems. multi-agent systems, pages 44–65, 2005.
  3. Harry Chen, Tim Finin, and Anupam Joshi. An ontology for context-aware pervasive computing environments. The Knowledge Engineering Review, 18(03):197–207, 2003.
  4. Marco Colombetti and Mario Verdicchio. An analysis of agent speech acts as institutional actions. In Proceedings of the first international joint conference on Autonomous agents and multiagent systems: part 3, pages 1157–1164. ACM, 2002.
  5. Richard O. Duda, David G. Stork, and Peter E. Hart. Pattern classification and scene analysis. Part 1, Pattern classification. Wiley, 2 edition, November 2000.
  6. Jean-Marie Favre. Foundations of meta-pyramids: languages vs. metamodels. In Episode II. Story of Thotus the Baboon, Procs. Dagstuhl Seminar, volume 4101, 2004.
  7. Maria Fox and Derek Long. PDDL2.1: An extension to PDDL for expressing temporal planning domains. Journal of Artifical Intelligence Research, 20:61–124, Decemeber 2003.
  8. C. Goddard and A. Wierzbicka. Semantic and lexical universals: Theory and empirical findings, volume 25. John Benjamins Publishing Company, 1994.
  9. Cliff Goddard. Cross-Linguistic Semantics, chapter Natural Semantic Metalanguage: The state of the art, pages 1–34. John Benjamins Publishing Company, 2008.
  10. Cliff Goddard. Semantic molecules and semantic complexityemoticon_unhappywith special reference to” environmental” molecules). Review of Cognitive Linguistics, 8(1):123–155, 2010.
  11. Bernardo Cuenca Grau, Ian Horrocks, Boris Motik, Bijan Parsia, Peter Patel-Schneider, and Ulrike Sattler. Owl 2: The next step for owl. Web Semantics: Science, Services and Agents on the World Wide Web, 6(4):309–322, 2008.
  12. Naresh Gupta and Dana S Nau. On the complexity of blocks-world planning. Artificial Intelligence, 56(2):223–254, 1992.
  13. Karen Henricksen and Jadwiga Indulska. Modelling and using imperfect context information. In Pervasive Computing and Communications Workshops, 2004. Proceedings of the Second IEEE Annual Conference on, pages 33–37. IEEE, 2004.
  14. Ian Horrocks, Peter F Patel-Schneider, Harold Boley, Said Tabet, Benjamin Grosof, Mike Dean, et al. Swrl: A semantic web rule language combining owl and ruleml. W3C Member submission, 21:79, 2004.
  15. Ora Lassila and Ralph R Swick. Resource description framework (rdf) model and syntax specification, 1999.
  16. Drew McDermott. The formal semantics of processes in pddl. In Proc. ICAPS Workshop on PDDL, pages 101–155. Citeseer, 2003.
  17. Deborah L McGuinness, Frank Van Harmelen, et al. Owl web ontology language overview. W3C recommendation, 10(2004-03):10, 2004.
  18. Igor Melcuk and Leo Wanner. Lexical functions and lexical inheritance for emotion lexemesˇ in german. Lexical functions in lexicography and natural language processing. Amsterdam/Philadelphia. John Benjamin, pages 209–278, 1996.
  19. Grundlagen der Theorie Unscharfer Mengen. Fuzzy decision support-systeme. 1994.
  20. Mazeiar Salehie and Ladan Tahvildari. Self-adaptive software: Landscape and research challenges. ACM Transactions on Autonomous and Adaptive Systems, 4(2):1–42, May 2009.
  21. Jeffrey Smith, Scott DeLoach, Mieczyslaw Kokar, and Ken Baclawski. Category theoretic approaches of representing precise uml semantics. In of the ECOOP Workshop on Defining Precise Semantics for UML, 2000.
  22. Robert J Sternberg. Beyond IQ: A triarchic theory of human intelligence. 1985.
  23. Giorgos Stoilos, Nikos Simou, Giorgos Stamou, and Stefanos Kollias. Uncertainty and the semantic web. Intelligent Systems, IEEE, 21(5):84–87, 2006.
  24. Wiebe van der Hoek. Epistemic logic for AI and computer science, volume 41. Cambridge University Press, 2004.
  25. Anna Wierzbicka. Semantics: Primes and Universals. Oxford University Press, 1996.
  26. Anna Wierzbicka. English: meaning and culture. 2006.
  27. Anna Wierzbicka. Mental lexicon. In The Slavic Languages: An international handbook of their history, their structure and their investigation. Mouton de Gruyter, Berlin, 2009.

[1] For further information the interested reader is also refereed to: http://lists.w3. org/Archives/Public/www-webont-wg/2001Dec/0169.html

[2] For further information the interested reader is also refereed to: http://www.cs.yale.


[3] see:


Copywrite © 2020 LOGICMOO (Unless otherwise credited in page)