User:Mike/About Swedish
Appearance
The work done on Swedish (and presumably the majority of languages) are very... should we say... inconsistent and spotty. There is no, nor has there ever been any, systematic work done on adding (Swedish) words, or to see what remains to be done. Hence, I'm thinking of doing a todo list, and try to figure out what should be done... Or at least making a list of "issues" which (may) need to be dealt with...
Templates
[edit]- The inflection/conjugation templates are a mess. Inconsistent names, partly incompatible uses (cases left "between the chairs"; cases covered by several templates). They make no use of the
{{infl}}
template - should they?- Adjectives:
- No proper template for those adjectives who never inflect for gender/number. Or should it simply be marked in the
{{infl}}
template? What would be the proper term? "uninflectable"? "unconjugatable"? - If a infl-template solution is chosen, which forms should be given there? For other forms, add a ====Conjugations==== header and put a table similar to the present ones thereunder. (Note: the display of the tables should be reworked into something which more resembles sv:wikt's - there are several ... issues with forms which probably are not quite correctly classified). Suggestions for the infl template:
- Quick inflection table (to be used as the header for a collapsible table?): Common predicative positive, (common predicative) comparative, (common) predicative superlative
- Absolute (no comparations): note not comparable
- Periphrastic: note comparative and superlative with mer and mest, respectively.
- Uninflectable: note no inflections (and no "Conjugations" header)
- Which templates (template names) should be used?
- Present set on en:wikt:
{{sv-adj-reg}}
,{{sv-adj-irreg}}
,{{sv-adj-abs}}
,{{sv-adj-peri}}
. - Present set on sv:wikt: (a set of 26 different templates..., basically
(-,-peri,-abs)x(-,-t,-ad,-dd,-el,-en,-er,-(v)d,-(k)d)
; exception is sv-adj-ad) - Suggestions:
{{sv-adj}}
,{{sv-adj-abs}}
,{{sv-adj-peri}}
, each with parameter-based switches such as{{sv-adj|(stem)}},{{sv-adj|(stem)|t}},{{sv-adj|(stem)|x|d}},{{sv-adj|(stem)|e|y}}
(technical point: special uses for x=a,d and for y=l,n,r). Further parameters: n= (for special neuter form spellings), alt= (for cases where the stem of the comparative/superlative differ from the positive). Possibly a no-parameter template{{sv-adj-none}}
. And finally, all templates should have a complete set of named parameters, for truly irregular adjectives (point in case: liten).- Pro: Fewer templates
- Con: A few words will require more complicated template calls.
{{sv-adj}}
,{{sv-adj-abs}}
,{{sv-adj-peri}}
{{sv-adj-ex}}
,{{sv-adj-abs-ex}}
,{{sv-adj-peri-ex}}
,{{sv-adj-ad}}
{{sv-adj-abs-ad}}
,{{sv-adj-peri-ad}}
incomplete- Pro: Easier use for the more complicated adjectives
- Con: More templates to create and remember the names of, and know when to use
- Present set on en:wikt:
- No proper template for those adjectives who never inflect for gender/number. Or should it simply be marked in the
- Adverbs
- Presently only for irregulars - a template for regular would be trivial to write.
- On the basis of standardization, perhaps also the non-adjective-derived should have a template (which simply would call
{{infl}}
with suitable parameters).
- Nouns
- Are "genitive" a proper 'case' in Swedish? When did we move to the present "enclitic" -s for genitive? See also [1] - someone's convinced 'genitive' would be completely wrong. Personally I'm not convinced as long as the formal use still seems to be to use what would literally translate into English as "Queen's of England men". (Though I agree it's vanishing from everyday use - but some still maintain that this is the correct way.)
- Addition of the form(s) which are used in compounds. (E.g. "varu-" from "vara", "kvinno-" from "kvinna").
- Already attempted: User:Mike/sv-noun.
- Doesn't add any genitive (attributes) (would it even be possible to base it upon the
{{infl}}
template *and* add a hideable section to it?) - May be a tad too complicated (both "technically" and "to use").
- Do however add the "compound form". (Note: a few words, e.g. "broder", have different compound form if they refer to singular or plural uses (broder- and brödra-, respectively. Not implemented.))
- Further, it doesn't address "nounal adjectives" such as laglös.
- Does not deal with plurale tantum/singulare tantum.
- Doesn't add any genitive (attributes) (would it even be possible to base it upon the
- Verbs
- Passive voice? (Not included in present templates). Should one add the "reciprocal" or "absolute" 'voices' too? Or should one, as has been suggested, simply add a "s-form" as a general catch-all? If so, how to mark which voices exist with a given verb?
- Special templates for compounded verbs (sammansatta verb); that is, verb which includes a particle. Sometimes non-intuitive inflections, in particular past participle which may be required to be used as one word instead of one or two as in other tenses.
- Do all sammansatta verb which can be written either as one word or two words actually mention this? (Cf. utge vs. ge ut in the sense of "publish")
- Attempted: User:Mike/sv-verb.
- Pros:
- Do use the infl template
- "inflection line" style, for greater consistency with other languages, esp. English
- Doesn't clutter the right hand side of the screen, which often(?) are busy with wp link box, images etc.
- Doesn't require a whole new ====Conjugations===== header/section.
- Cons:
- Since it uses the infl. template, it requires one ====Verb==== header for each variant - perhaps not as important for verbs as for nouns?
- For the sake of consistency, it'd require noun, adjective and other PoS templates to be equivalently reworked.
- The collabsible inflection-line template is something quite different from how other languages do it.
- It does not easily lend itself to any future addition of a "historic forms" section, without making the template very large...
- Suggestion: Rework it into a simpler "active voice" only template (just as the present ones are already), and let the passive forms, and the historic forms, show up under a ====Conjugations==== header: more to write, sure, but most users would probably be content to have the active voice anyway, until someone can add the full setting.
- Pros:
- Adjectives:
- The "infl"-templates 'should' only show the most common forms, with extra variants in 'another conjugations template'. Passives/"s-forms" and genitive forms would probably be the forms which fits the description of a form which shouldn't be present in the main infl-table. But would it be possible/desirable to add them to a hideable section of the infl-table? IMO they are too few (in contrast to the French verbs) to really motivate a whole section of their own... Note that this is for "Contemporary Swedish"; see below about historic inflection patterns.
- "form of" templates. Which exist already, and which are needed by/suitable for Swedish? Which new needs to be added?
- Bot(s) to add "form of" articles.
- Templates for historic inflectional patterns; see below.
- Once a new set of templates are done, one would need a list of articles to be updated.
Categories
[edit]- Which are around?
- Which should be around?
- Which articles are categorized properly? At all?
Headwords, inclusion
[edit]- Which words should we add? As many words which in English would be a SoP phrase, in Swedish is one single compound word, should they be considered as a proper word, or a SoP?
- Which words are already present? That is, which *Swedish* words are there? Linklists only tell if the page exists, not if a ==Swedish== header exists... The same about individual senses.
- Todo: More lists of words used in Swedish. User:Mike/Word lists are insufficient/impractical; the lists would also need frequency indications. Pick words from sv:wp?
- Example sentences. A database of concordances from sv:wp? Would that be suitable to put on (en:)wikt? How to make sure it could be used also on other wiktionaries?
- Possibility to - from database dumps - check up on which words lack one or more pieces of important information? (Etymology, Pronunciation, inflection [table], category, ...). How and where to publish such lists?
"Request for addition of senses". How, and where, is a user supposed to be able to request a missing sense? (N.b. of course not only about Swedish). Anons has the "Feedback" system; logged-in users not.This is done by {{defn}}.- Reflexive verbs: On the page for the main verb with a tag {{reflexive}}, or on their own pages? Tags:
- Pro:
- Intuitive, don't have to bother with finding the right pronoun when searching for existing words
- Tells the user straight away that it is a reflexive verb - s/he may have found it in a text and missed that pronoun/never understood it belonged to the verb
- Con:
- How to deal with phrasal reflexive verbs? E.g. "ta sig till" and "ta till sig" are quite different, but as we usually put verbs such as "ta till" on their own page, the verb "ta sig till" doesn't easily fit at neither "ta" nor "ta till" (and a reflexive marker on "ta till" would yield the reflexive phrasal "ta till sig" instead...). Or should one apply another set of rules for verbs with the reflexive pronoun in the middle and give those, and only those, headwords on their own?
- Pro:
- Proper nouns, in particular place names. As far as I understand, attributive use is (presently) a requirement for inclusion. Is that to be interpreted as to allow also place names which are used in compounds? Should then the entry be located at the base form of the word (e.g. "Stockholm") or including the joining -s- (e.g. "Stockholms-")? It would then, IMO, not even be clear which PoS to give this. Neither whether it should be given a capital letter or not - one could conceivably come up with about as many compounds where "Stockholms-" should be capitalized as not, though they would nevertheless have exactly the same definition. (*The* Swedish dictionary, SAOB, the net edition, sidesteps this by using "STOCKHOLMS-" as its entry.)
- Needs to be structured up.
- As new templates are created, they need to be added
- Old, discontinued templates should be removed from the page and deleted.
- Better design in general would be needed - e.g. the examples of the form-of-templates' use are pretty much useless atm.
- Are they updated? Contain what needs to be there, but not overly large?
- Should there be both a "basic" set and an "advanced" set?
Historic variants
[edit]- How to mark words used before the last spelling reform of early 20th century? Should they be linked from the modern spelling, and if so, how many, and how old, variants should be mentioned? Is "Alternative spellings" even a good header for this? Which (context) tags should be used to mark [the old spelling] when it was used? (Old Swedish, Middle Swedish, Modern Swedish?? [The Swedish of today is called "Contemporary Swedish" (sv: "nusvenska", lit. "now-Swedish")...]) Would there be any point at all in using the "archaic" tag if these three tags were to be used? Is "älfen" an 'alternative spelling' of "älven", or an inflection of 'älf' (which in turn is an 'alternative spelling' of "älv")?
- Inflections of various historic variants: at least Old Swedish nouns used a larger set of inflections (rather: "cases") than more modern Swedish does (hence new templates needed for those variants) and verbs need a *large* inflection table for Swedish up to and including "Modern Swedish" (which is to include inflection for person); only "Contemporary Swedish" has cut those forms; should those extended tables get a ====Conjugations==== header, or could they be part of the hidden display? If a word has the same spelling today as then (often it will), should there be two different templates used; one for the modern pattern and one for the archaic, or should they be merged into one which has a "modern" section and one "archaic" section? If so, how does the user who is unfamiliar with the old inflections add that template?
Headers
[edit]- Related terms:
- Are "compounds" really a form of "Related term"? Phrases in which the words is included?
- Under which PoS to put "Related terms" if the related terms are related to several PoS's?
- Alternative spellings
- Used as a L3 header; the same as "Etymology" and "Pronunciation". What about words where two senses has the same etymology but only one of them has an alternative spelling? (todo: example thereof) Of course there may also be cases where "Pronunciation" differs between two senses of the same etymology... See also the issue under "Historic variants".
Rhymes
[edit]- Some add rhymes to Swedish words/lists of rhymes in Swedish to en:wikt; should such lists be here, or would it make more sense to restrict such lists to sv:wikt? Same of course for other FLs.
Pronunciation
[edit]- There is a bot which has added audio to Swedish words; sometimes it fails to differentiate between homographs. Those articles needs to be checked; do we have a list of Swedish words it has added pronunciation to? Often those (problematic) articles also need to be given separate ===Etymology=== headers too.
- Another issue with the audio files obtained from the shtooka project is that the recordings for nouns tend to incorporate the indefinite article too... So, the audio file for "tack" *really* is a recording for "ett tack".
PoS
[edit]- Are the participles best classified as verb (forms) or as adjectives? They (well, only the past participles [or should that even be past participle, or should it be perfect participle?]) inflect as if they were adjectives. Usually, we don't add inflection information to inflected forms, so there may very well be resistance to marking them as a "verb form" *and* give them a - basically - adjective based inflection table.
Etymology
[edit]- Very few etymologies added.
Sources/References
[edit]- SAOB (Svenska Akademiens OrdBok) are commonly referenced through links of the form "http://g3.spraakdata.gu.se/saob/show.phtml?filenr=*/**/****.html" where all "*" stand for one digit. These links are *not* stable. They change every now and then in the sense that a link that today links to word <foo>, tomorrow may link to <foob>. Links of the form http://g3.spraakdata.gu.se/cgi-bin/osa/indexsok.cgi?string=BAR seems to be more stable, though the drawback is that they only end up at a list of search results, not in the actual entry.
- This I should study at some point. Or whoever decides to take up the slack and get to work with the verb conjugation templates!
...to be continued...
...(if you see more issues which needs to be addressed, please add them here! Or on my talk. Or...)