User Tools

Site Tools


pos_marking

Different auxlangs feature part-of-speech marking to different degrees. When it comes to Folkspraak, there is perhaps only one place where the issue really matters - in distinguishing between certain noun+verb pairs, generally when the noun is singular or the verb is in the present tense or imperative mood.

Where there is a root noun, there will often be an obvious associated verb, and where there is a root verb, there will always be a family of associated nouns. If the associated word can be produced by means of an affix or compounding, then there is no problem.

But sometimes, the noun that's related to the verb isn't covered by any of the pre-existing affixes, and can often be vaguely defined as “that which is used in the action of the verb”. One example is “song/sing”, where the closest affix for deriving the noun is the past-participle ending - but that usage would be rather dodgy. Other examples include “drink/drink”, “blood/bleed”, “deed/do”, “bath/bathe”, and “belief/believe”.

The big question is this: how should these words be morphologically related to one another? Here are some options.

  • Use the same word for the noun and verb, like with “drink”. This is simple, and the fact that nouns often have an article helps. The disadvantage is that ambiguity can arise due to the lack of marking - hence the present page in the wiki!
  • Inflect all verbs for the present tense and imperative mood, e.g. with -e or -a. This would make the noun/verb distinction always clear, and derivation between the two would be simple. The disadvantage is all verbs having an extra syllable, which would often feel tedious and redundant.
  • Have markers for derived basic nouns/verbs. Root nouns and root verbs would have no markers, but their related derivatives would. This offers similar advantages and disadvantages to the verb-inflection system, but it's hard to imagine what the noun marker would be.
  • Have separate roots, like with “song/sing”. This gives you clarity, but it makes the language harder to learn - although if the number of such pairs was low, the increased learning burden might not be so bad.
  • Combining the previous options, have a system for derivation between roots, like the “vowel triangle” in the abandoned Afrihili project. This could work, but it would be a bit weird, and it may prove very difficult to establish rules, given the lack of patterns in ablauting across the Germanic languages. And it could create a silly situation of runaway derivation, with every word in the language being interpreted as having a twin.

The POS-marking issue relates to the language as a whole, and should be borne in mind when the noun, adjective, and verb morphologies are being defined.

Discussion

I prefer to use as much word-derivation and -composition as possible. Possible solutions for the mentioned examples:

  • sing/song: singe as the verb and singung as the noun
  • do/deed: done as the verb and donung as the noun
  • belief/believe: true as the verb and trurid as the noun
  • bleed/blood: blud as the noun and lose blud as the verb
  • bath/bathe: bade as the verb and bad-sted as the noun
  • drink/drink: (no satisfying solution yet, maybe drink as the noun and take en/de drink as the verb or a new affix for these situations)

Though I admit that this ain't possible or satisfactory in all cases. ~~~ Fenris

Affix overloading

This is a continuation of the POS-marking issue. In summary, the same affix can be used on words from different parts of speech to produce words of different meanings. Here are three examples.

  • Suffix -e for plural AND infinitive; thus noun/adjective + E = plural, but verb root + E = infinitive. If we have full POS-marking, this works nicely, because ambiguity is impossible, although to people who don't know the words there would be apparent ambiguity. If POS-marking is limited, however, then ambiguity becomes possible, and this can be problematic. It's not too bad with, for example, “I want drinks” versus “I want to drink”, but in the case of “I want to wound” versus “I want wounds”, we're in trouble.
  • Suffix -er for agents AND positive comparative adjectives. The rule would be that noun/verb + ER = person, but adjective + ER = “more …”. It is likely that many roots will end ER (offer, bruder, fujer, etc.) so whatever we choose, there will be superficial ambiguity here. The only restriction with this overload is that when a standalone adjective is to denote a person, it cannot take the personal suffix; e.g. “the good one” must be de gud and NOT de guder. And an unfortunate consequence is a “stuttering effect” in certain derivatives, e.g. “offerers” being offerere.
  • Past tense and past/passive participle the same. As everyone in the world knows (well, 10% of 'em) in regular weak English verbs, the same word is used here, ending “-ed”. This could theoretically happen in Folkspraak too, although most proposals seem to insist on a distinction.
pos_marking.txt · Last modified: 2014/10/16 14:06 by ob