I'm trying to create an educational system, and I would like to have more taxonomies than a simple part of speech. For example, I would like to be able to categorize words into:

  1. Tree => object
  2. Kindness => attitude
  3. Keyboard => computer part
  4. Car => transportation, object

A better example can be found at this English word groups page.

Do we have such classification in any corpus of English language? Is it related to linguistics at all, to categorize words by other taxonomies too.

  • I think you are looking for ontology, not taxonomy. And, yes, there are several. – prash Aug 29 '15 at 16:02
  • 1
    It's very limiting to stick to categorizing single words. For linguistic analysis, we need phrases. We have always needed them -- read over Aristotle's Categories, where he considers what expressions can answer a given question. – Greg Lee Aug 29 '15 at 16:32
  • @parsh, ontology means studying existence of things. How that can be used as the notion of classification and categorization? – Saeed Neamati Aug 29 '15 at 17:01
  • It's not entirely clear what you're looking for. You could be interested in subset relations between things referred to by any linguistic expression; or you could be specifically interested in conceptual relations expressed by lexical expressions (words). In the latter case, you would be looking at e.g. "reindeer" and "mammal" or "herbivore", but not "white reindeer" (but: some languages have a lexical item for "white reindeer"). BTW ontology relates to classification because a classification is of things that exist. – user6726 Aug 29 '15 at 17:42
  • 1
    No, you're talking about ontology in philosophy. I'm talking about ontology in linguistics and computer science. en.wikipedia.org/wiki/Ontology_(information_science). See mpi-inf.mpg.de/departments/databases-and-information-systems/…, for example. – prash Aug 29 '15 at 18:19

Since you asked whether there is a corpus with such a classification, I'm going to give you an example of a semantic tagger. The UCREL Semantic Analysis System (USAS) can be run on a corpus just like a parts of speech tagger and relies on a taxonomy of 21 general levels of classification:

enter image description here

Below that there are up to four sub-divisions, here is the classification for time expressions:

enter image description here

There is a free web-service where you can try out the semantic tagger.

Here is an example of the sentence "My dog loves biting the postman." It's a bit cryptic, but that's because it's supposed to be machine-readable.

My_Z8 dog_L2mfn loves_E2+ biting_B1 the_Z5 postman_Q1.2/I3.2/S2.2m ._PUNC
  • 1
    that's very interesting. It's getting close to what I have in mind. Does it have help on reading the result and understanding it? – Saeed Neamati Aug 30 '15 at 8:20
  • Yes, you need to look up the codes in the classification tree. For exampe, 'loves_E2+' is something to do with "emotion" since it's in category E. – robert Aug 30 '15 at 11:46

Another kind of resource of interest is a wordnet: It shows several relations between words like synonym, hyperonym, hyponym. The prototypical wordnet is the Princeton wordnet available from http://wordnet.princeton.edu/ .

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.