User:Dan Polansky/Morphology and etymology

What follows are notes on treatment options for morphology and etymology. There is Etymology.

Templates
Key templates are der, af, surf and cog. There is com. There used to be etyl, used with m.

Czech compounds
See also Appendix:Compounds.


 * černohorský = černý + -o- + hora + -ský.
 * An alternative would be černý + -o- + horský.
 * A non-Occam's-razor alternative would be černo- + horský, where černo- would be defined as a combining form with etymology černý + -o- or čern- + -o-. We generally don't do that. We would need combining forms for great many nouns and adjectives. Treatment of other Slavic languages can be seen by navigating from the translation table at chernozem, and seems to match the minimalist approach. German compounds follow similar minimalist treatment, using -s-, -n-, etc.
 * modrooký = modrý + -o- + oko + -ý.
 * mrakodrap = mrak + -o- + drápat. Alternatively, one could mark the stem of drápat as drap. What makes little sense is to posit suffix -drap.
 * zlatokop = zlato + kopat. The -o- is skipped since zlato already ends in o. No suffix -kop.
 * kočkopes = kočka + -o- + pes.
 * Mrakoplaš = mrak -o- + plašit. A translator's invention. Here, again, only the stem of plašit is used.

English neo-classical compounds
See also Appendix:Compounds.


 * biology = from German, surface analysis bio- + -logy
 * speedometer = speed + -o- + -meter. An alternative would be speed + -ometer, which I disprefer since WT:MW and WT:OED mostly do not use this analysis except for a handful of suffixes and since, contrary to Occam's razor, it posits existence of entities whose existence is not obvious and that are not necessary for analysis, namely all those -oX putative suffixes in addition to -X suffixes.
 * A related RFD will be archived at Talk:-otomy.
 * -osis would be -o- + -sis, rather non-traditional. It may be a special case.
 * -itis would be -i- + -tis, rather non-traditional. For -itis, one may note -ίτης and its note about reanalysis of πολῑ́-της as πολ-ῑ́της, so it may be a special case.
 * morphology = morpho- + -logy. It could be analyzed as morph- + -ology (odd in my view) or morph- + -o- + -logy (interesting Occamian minimalism doing away with morpho-), but we don't do that. Doing away with morpho- and only keeping morph- would be interesting but rather non-traditional. The current approach is to often duplicate prefixes X- as Xo-, e.g. arthr- and arthro-. A listing of such pairs is at User:Quercus solaris/Remedial stuff and Appendix:Medical prefixes, suffixes, and combining forms. There is List of medical roots, suffixes and prefixes, which handles the duplication by means of e.g. "arteri(o)-", and does not associate -o- to the right in its examples.

Names for -o-: "interfix" and "thematic vowel" per encyclopedia.com. "linking element" and "linking morpheme". Other names are in the non-academic taalportaal.org. MW calls -o- an "affix", OED a "connective", AHD does not call it anything but says it is "Used as a connective".

These compounds are currently not classified as compounds. See also Category:English compound terms.

Original research in etymologies
There does not seem to be a vote forbidding original research in etymologies. A discussion:

Some original research may be hard to avoid. For instance, dictionaries do not usually explicitly provide an analysis like modrooký = modrý + -o- + oko + -ý; we would have to omit that morphological analysis altogether.

Original research in reconstructions
This is covered by a vote: Votes/2013-10/Reconstructions need references.

Surface analysis
Surface analysis, whether it should be called that way, seems useful. See surf. We often say "Equivalent to X" without a template. I sometimes used "As if X", which is shorter.

Surface analysis is often morphological analysis. Take Czech listopad: without surface analysis, we would need to trace it to Proto-Slavic *listopadъ and be done with it. But we may note it is in fact list + -o- + pad- (from padat), which is a plausible morphological analysis. By tracing items only to Proto-Slavic, we would have to skip the morphology.

For neo-classical compounds, surface analysis fills categories for prefixes and suffixes. Prefixes and suffixes are not only etymological but also morphological categories, and do not need to be productive.

Productiveness of affixes
We mark some affixes or their senses as "non-productive". Examples: in- and con-. -ion is completely marked as non-productive and is in other dictionaries per OneLook.

Forbidding non-productive affix entries would greatly limit morphological analysis, especially in a hybrid language such as English. There would be no way to list terms suffixed with -ion.