Morphology Paradigms

[Stephen Carlson]

I split this fascinating topic into its own thread.

[RandallButh]

Perhaps a different way of approaching what I think Peter was saying is - when do we agree with what grammars (and lexicons and commentaries) have always told us? What methodology do we have to validate or invalidate their assertions? (cf. John Lee's article that was shared here recently on lexicons, their dependence on earlier lexicons, and continuing to promulgate old, invalid assumptions)

If something like these odd feminine 2nd declension nouns and 2 termination adjectives 'smells funny' - what is our means of verifying what we've always been taught? To prove or disprove Peter's assertions we'd really want to be able to do exhaustive searches on feminine 2nd declension nouns with adjectives.

Paul, categories like these exist in most every language I know. Buenos dias ! (Please note the concord, masculine with an -a class noun [like señorita].)
Arabic nouns are typically tri-partite in the singular (nom, gen, acc) but binary in the plural (nom, oblique) some adjective forms are binary, etc.
Not everything in a language gets leveled when new foci of regularization crystalize, and the other forms may stand out as "irregular." The best approach is to learn the language to a functionally communicative level and these things become ignored curiosities.

Imagine English: we may OVERLOOK an item, but we better not do that when we are supposed to OVERSEE something. Of course, an OVERSIGHT may get us assigned to guard duty at some OVERLOOK on the city dump.

Languages are funny.

[Barry Hofstetter]

Paul, categories like these exist in most every language I know. Buenos dias ! (Please note the concord, masculine with an -a class noun [like señorita].)
Arabic nouns are typically tri-partite in the singular (nom, gen, acc) but binary in the plural (nom, oblique) some adjective forms are binary, etc.
Not everything in a language gets leveled when new foci of regularization crystalize, and the other forms may stand out as "irregular." The best approach is to learn the language to a functionally communicative level and these things become ignored curiosities.

Imagine English: we may OVERLOOK an item, but we better not do that when we are supposed to OVERSEE something. Of course, an OVERSIGHT may get us assigned to guard duty at some OVERLOOK on the city dump.

Languages are funny.

There's actually an explanation for dia in Spanish. It descends from the Latin 5th declension noun dies, which in most contexts was treated as masculine, except when a specific day was in mind (diē constitūtā, on the appointed day). Morphologically as Latin changed into Spanish it became an a-class noun, but retained its traditional gender nonetheless.

Of course we understand your illustration, but we had better not stand under the overlook when overseeing it.

I recently told my 7th grade Latin class the etymology of nice. As expected, they immediately began "complimenting" one another on how "nice" they were...

Languages are funny, and we should be able to have fun with them now and again.

[paorear]

Languages are funny.

Languages are funny, and we should be able to have fun with them now and again.

I completely agree and hope that I didn't cause too much consternation with re-opening this thread. Randall knows me better than Barry, and is a little more aware of my background.

I'm chiefly most interested in making informed observations on language, and particularly in Koine Greek in this case. And that informed-ness can certainly come primarily from experience in simply reading the language and spending many years with it. (Which I've also done.)

However, I also have enough background as a software engineer (as well as linguistics and biblical Greek) that I like to see data. And I wish to be open enough to consider alternative ways of asking a question of either the text or the language and to consider alternative ways of defining a solution.

That was my primary interest in entertaining Peter's assertions - was he seeing something we weren't, however much I might have doubted it? If in looking at the data there might be alternative explanations, how might that change things? In Peter's case, how about something like - 2nd declension feminine forms are really masculine and always seem to take a feminine article? (I'm not saying this is so.) Again, the data would be the most interesting thing to look at and review what is happening with modifier gender agreement (or lack thereof). And certainly, as Randall points out - languages commonly have this "untidy-ness". But before we settle on that, let's see the data.

On a wholly unrelated matter this morning, my daughter and I were discussing how it's better to have strongly held opinions, but held openly, then simply strongly held opinions. I hope to strive for the former as I believe the rest of you do as well.

[Jonathan Robie]

In terms of producing "a list of nouns and adjectives..." - for the purposes of this original thread, a listing of all 2nd declension *feminine* nominals in verses with adjective modifiers of any gender I think would help put a nail in this particular coffin.

Ah, so there's a syntax question here. I'd hoped I could help just by providing inflectional classes but it sounds like we need to extract the constructions from a treebank.

Yes, but I already have the treebanks.

There are an awful lot of queries where having a list of words would make it easier to leverage the treebank well, e.g. a list of words of a given inflectional class, a list of verbs of knowing, expression, and perception, etc.

And some of these lists could be done by people who aren't programmers.

[Barry Hofstetter]

That was my primary interest in entertaining Peter's assertions - was he seeing something we weren't, however much I might have doubted it? If in looking at the data there might be alternative explanations, how might that change things? In Peter's case, how about something like - 2nd declension feminine forms are really masculine and always seem to take a feminine article? (I'm not saying this is so.) Again, the data would be the most interesting thing to look at and review what is happening with modifier gender agreement (or lack thereof). And certainly, as Randall points out - languages commonly have this "untidy-ness". But before we settle on that, let's see the data.

That would be similar to saying "Black swans are really white swans that happen to look black, so in actuality there are no black swans." If a second declension form is really masculine but takes the feminine article, then how would we distinguish masculine from feminine? If the noun is regularly found with the feminine article, then it is a second declension feminine, as rare as those happen to be. Thinking outside the box is one thing. Tearing up the box, putting it through a micro-cut shredder and then burning the remains is quite another.

[Jonathan Robie]

That was my primary interest in entertaining Peter's assertions - was he seeing something we weren't, however much I might have doubted it? If in looking at the data there might be alternative explanations, how might that change things? In Peter's case, how about something like - 2nd declension feminine forms are really masculine and always seem to take a feminine article? (I'm not saying this is so.) Again, the data would be the most interesting thing to look at and review what is happening with modifier gender agreement (or lack thereof). And certainly, as Randall points out - languages commonly have this "untidy-ness". But before we settle on that, let's see the data.

That would be similar to saying "Black swans are really white swans that happen to look black, so in actuality there are no black swans." If a second declension form is really masculine but takes the feminine article, then how would we distinguish masculine from feminine? If the noun is regularly found with the feminine article, then it is a second declension feminine, as rare as those happen to be. Thinking outside the box is one thing. Tearing up the box, putting it through a micro-cut shredder and then burning the remains is quite another.

But teaching what we've learned about the box is a third thing. With the data Paul is asking for, we can show what the pattern is, and someone like Peter can look for counterexamples, we can look at proposed counterexamples to see if they actually exist, etc.

I think we can really improve our teaching by making it easy to generate lists of examples that show how the language works. Most of the time that means explaining what existing grammars mean. Sometimes we may say things existing grammars do not, e.g. in some of the threads on word order. Teaching and research are not miles apart.

It's fair for a student to ask for examples that illustrate what we are teaching.

[MAubrey]

Here's the current version of my own analysis of Adjective morphology paradigms, for what it's worth:

https://docs.google.com/spreadsheets/d/ ... sp=sharing

As I see it, there are basically two types of binary gender (neuter vs. non-neuter) adjectives, those with dative plurals formed with -οις and those formed with -σιν.

[Barry Hofstetter]

But teaching what we've learned about the box is a third thing. With the data Paul is asking for, we can show what the pattern is, and someone like Peter can look for counterexamples, we can look at proposed counterexamples to see if they actually exist, etc.

I think we can really improve our teaching by making it easy to generate lists of examples that show how the language works. Most of the time that means explaining what existing grammars mean. Sometimes we may say things existing grammars do not, e.g. in some of the threads on word order. Teaching and research are not miles apart.

It's fair for a student to ask for examples that illustrate what we are teaching.

Grammars and lexicons do this quite nicely.

[Jonathan Robie]

But teaching what we've learned about the box is a third thing. With the data Paul is asking for, we can show what the pattern is, and someone like Peter can look for counterexamples, we can look at proposed counterexamples to see if they actually exist, etc.

I think we can really improve our teaching by making it easy to generate lists of examples that show how the language works. Most of the time that means explaining what existing grammars mean. Sometimes we may say things existing grammars do not, e.g. in some of the threads on word order. Teaching and research are not miles apart.

It's fair for a student to ask for examples that illustrate what we are teaching.

Grammars and lexicons do this quite nicely.

Often. So why would you want this in digital form?

• For the same reason we would do this on paper - recording observations and verifying them is important.
• If observations are in digital form, we can derive new insights that the original observer may not have been thinking about by querying their observations.
• I'm not convinced that we do a good job of teaching some aspects of the language, and we can use digital representations to explore the data and add to what we can describe in the grammars.
• Whoever wrote the lexicons and grammars did precisely this kind of observation, but we generally have access only to their summaries. I would LOVE to have their detailed observations, the things they wrote down on cards and such, in digital form. As far as I know, most of this is lost.
• Neither paper nor digital observations are perfect. But digital observations are more easily compared to each other so we can see where they agree or disagree. For instance, I can compare PROEIL's analysis to GBI's analysis of the same sentence, or compare query results across them. Or I can sit in an SBL presentation and call up examples of the constructs that are being discussed.

Of course, this is just one way of knowing about Greek. People like you who read enormous amounts of Greek are indispensable, that experience can't be replaced by digital resources. Ideally, digital resources contain observations made by people like you.