http://micra.com/COSMO/COSMO.owl
COSMO ontology, version 0.90-1751 by Patrick Cassidy cassidy@micra.com last edit 20210108: 1 AM
Intended as an OWL-format ontology of primitive and common concepts suitable for use
as an interlingua for semantic interoperability or language understanding.
Includes representations of common concepts learned by children at an early age, plus
the most frequently used words of English..
Has representations of more than 12,000 of the most frequently occurring words
in text corpora, with references to WordNet and
OpenCYC categories having (approximately) the same meaning.
As of Rev1743: 23647 owl:Classes
TBD:
22,198 distinct Single-word WordNet labels
23,961 Single-word labels including also en and ldoce labels not in WordNet
18,642 distinct WordNet synsets linked (some appear more than once)
Current phase of development : Current phase is to add in 4556 Words in the BYU 20,000 most frequent not already in the COSMO:
Finished adding in 519 words from the most common 10,000 words in the BYU most-frequent list
not already in COSMO: rev 1727 20200915
Finished adding in 438 words from the most common 8000 words in the BYU most-frequent list
not already in COSMO: rev 1707 20200718
Finished adding in 1720 words from an American Sign Language on-line
dictionary not already among the 'wordnet'. or 'en' tags, nor in
the openCYC nor BNC lists. rev 1689 20200409
Finished adding in representations of about 1400 common words in OpenCYC
not already represented in earlier versions at rev1623.
Finished adding in representations of the 6300 most frequent words in the BNC corpus
at rev 1577.
Brief history:
Effort at finding the full set of semantic primitives started with
representation of the'defining vocabulary' of 2130 words in the
Longman Dictionary of Contemporary English. (LDOCE).
The Longman-LDOCE vocabulary words and their WordNet mappings were
completed to end of 'Z' words (mapping completed) At rev1093. Started adding
WN labels to entries that have ID's the same as common words, but which
were not labeled (1600 of them at rev 1139) 20160423. Finished 'Z' words.
Added additional 900 words in the vocabulary of a 6-year old (from 1903) at rev 1283
20171209
Has 19 top-level classes under 'Thing'
The representation of English words was expanded by adding in words from the 5000 most
frequent words in the BYU corpus, and words from the 6300 most frequent word
in the BNC corpus (that were not in the BYU list). Finished at rev1577 20190729.
Protege 3.5 Metrics for rev1751f (20210108: 1 AM) ca. 24059 OWL Classes , ca 44 of which are synonyms ca. 1357 relations ('properties' not including 84 datatype) Ca. 21216 restrictions [ca. 284 metatypes (rfds:Class) ] *** Consistency Check: Prot 3.5 Fact++ for rev1751f (20210108: 1 AM) on XPS15 4661 sec OK Prot 3.5 Pellet for rev1751f (20210108: 1 AM) on XPS15 3511 sec OK Prot 5.2 Fact++ rev1751f (20210108: 1 AM) on Toshiba 8731 sec OK
Protege 5.2 Metrics for rev1735f (20201016: 11 PM) 23542 Classes (includes metaClasses) 1353 object properties 83 data properties 57667 'subclass' axioms 255 'disjoint' axioms 80 'Equivalent Classes' axioms 44 'Annotation Properties' 634 'SubObjectPropertyOf' axioms 344 'InverseObject Properties'
Uses elements of the OpenCyc OWL version 0.78, SUMO,
BFO and DOLCE ontologies, as well as elements created specifically for COSMO.
Added ontology elements corresponding to common English words (among the 5000 most
frequent in the BYU corpus). Continuing with addition of representations of
other common English words from an American Sign Language dictionary..
Parts of COSMO were closely aligned with elements of
other public ontologies such as OpenCyc, SUMO, and BFO,in order to ease the
translation of COSMO ontologies to those other structures, but the
conceptual structure of COSMO (hierarchy and relations) is too different from
those other ontologies to allow simplistic automatic translation; it is
anticipated that translation will be possible, but will require 'bridging
axioms' to convert the terms and syntax of COSMO to those of any other
ontology, and vice-versa. The meanings of the COSMO elements must be interpreted
only from the structure of the COSMO ontology and from the full COSMO documentation.
**NOTE** According to the documentation of those other ontologies,
they are **freely usable** by the public, though they remain copyrighted
by their originators and appear to require only acknowledgment of their
use (more detail below). No copyright restrictions are attached to
materials added in the COSMO project, therefore the only copyright
restrictions for use of this ontology are those placed by
the developers of the OpenCyc, SUMO, and BFO on parts derived
directly from those works. Those derived parts will include some
of the labels, and parts of the documentation, but the hierarchy and relations
are those **specifically created** for COSMO, regardless of whether documentation
or names were adopted from other ontologies. If necessary, the .
documentation text or names can be changed to avoid copyright issues.
****************** Source Materials ****************************
***** WordNet, OpenCYC, SUMO, Random House Webster **************
*******************************************************************
Relation of COSMO to other ontologies:
The COSMO ontology has a structure and basic viewpoint that
differs in some significant parts from that of the ontologies from which it
has derived materials, and the main parts of the hierarchical
structure and relations are not primarily derived from any of
the referenced ontologies, though some of the hierarchy is included because
of its use in WordNet. Most basically, the representations
were intended to adhere as closely as possible to **linguistic intuitions**
about the meaning and usage of English terms, while specifying
the meanings in a logically precise manner. Every element added to COSMO
is individually evaluated for its utility and validity within the conceptual
structure of the COSMO ontology, and is not derived or adopted solely or
mainly on the basis of the appearance of a similar concept in another ontology.
Certain individual subtype relations are similar to those in OpenCyc or SUMO;
but because the basic hierarchical structure of COSMO differs from the
other ontologies, logical inference using these relations will arrive at
conclusions that cannot be aligned directly with either OpenCyc
or SUMO. No simplistic mapping between COSMO and these other ontologies is likely to
enable accurate inference. The documentation derived from OpenCyc and SUMO
is provided as a means to **reference** similar concepts in
those other ontologies, and to explain similarities and differences, for
the convenience of those who are familiar with those ontologies.
A few of the entres were included solely because of their presence in OpenCYC,
but these are usually specific to CYC and are icluded solely to permit alignment with
CYC concepts for those who are using the full CYC ontology. These are
not essential to the structure of the ontology and could be removed.
Contents derived from OpenCyc and SUMO are copyrighted and
made freely available for public use under the terms found
in the documentation for those works (see below). Materials
added specifically for the COSMO project are not copyrighted.
The contents derived from SUMO are copyrighted by the IEEE and
made freely available for public use. For more detail see:
see http://www.ontologyportal.org
A description of the SUMO project can be found in:
Niles, I., and Pease, A. 2001. Towards a Standard Upper Ontology. In
Proceedings of the 2nd International Conference on Formal Ontology in
Information Systems (FOIS-2001), Chris Welty and Barry Smith, eds,
Ogunquit, Maine, October 17-19, 2001.
The Contents of the OpenCyc OWL version used in this project are found at:
http://www.cyc.com/2004/06/04/cyc
OpenCyc materials are copyrighted and licensed for free public use
under the GNU 'LGPL' license. The OpenCyc documentation reads:
************ OpenCyc copyright notice ******************
Copyright Information OpenCyc Knowledge
Base Copyright 2001-2004 Cycorp, Inc., Austin, TX, USA. All
rights reserved. OpenCyc Knowledge Server Copyright
2001-2004 Cycorp, Inc., Austin, TX, USA. All rights
reserved. Other copyrights may be found in various files.
The OpenCyc Knowledge Base The OpenCyc Knowledge Base
consists of code, written in the declarative language CycL,
that represents or supports the representation of facts and
rules pertaining to consensus reality. OpenCyc is licensed
using the GNU Lesser General Public License, whose text can
also be found on this volume. The OpenCyc CycL code base is
the "library" referred to in the LGPL license. The
terms of this license equally apply to renamings and other
logically equivalent reformulations of the Knowledge Base
(or portions thereof) in any natural or formal language.
See http://www.opencyc.org for more information.
************ OpenCyc copyright notice ******************
************ Random House Webster Electronic Dictionary ******************
"RHW": Definitions described as coming from the 'Random House Webster' (RHW)
refer to the Electronic Dictionary 'Random House Webster's
Unabridged Dictionary' on CD (2002) from Random House, Inc.
and Multimedia 2000 Inc.(a paperback version is still available
at: http://www.randomhouse.com/category/reference/)
OTher dictionaries:
"MWD": The Merriam-Wester Dictionary CD-ROM version
"MWCD" The Merriam-Wester Collegiate Dictionary (print version)
"RHUD" or "RHUDEL" The Random House Dictionary of the English Language (Print version,
second edition, unabridged, 1987)
Most of the entries have annotation references to WordNet
('wordnet' and 'wnsense' relations). The WordNet version referenced
is WordNet 2.1 (see http://wordnet.princeton.edu/).
Because the WordNet hierarchy differs in a number of places
from that of COSMO, these pointers are only informative, and may not
be useful for accurate automatic conversion of WordNet sense tags (synsets) to
the corresponding senses in COSMO, but at least *part* of those WordNet word
senses that are referenced will correspond to the intended meaning of the
COSMO ontology element, and those words will, in some context, refer to
the intended COSMO sense.