This is likely a long shot, but I am going to try anyway. I didn't know anywhere else to turn, but I am hoping for some real and meaningful engagement in this thread.

The Short Story: A while back, I developed a test program to programmatically generate new words for Ngwa (it also works for Igbo in general, but that's neither here nor there). The program was not as effective as I would like for it to be, so I scrapped it. I restarted my efforts about four months ago, but from a new angle---using dependency parsing.

It turns out dependency parsing is not easy for Ngwa (and consequently, Igbo). I have had to rewrite this parser twice and redo my corpus at least three times over the course of the last four months. What I've learned from all of this is that on a technical level, the language is complex (obviously), but far more complex that Igbo language researchers have given it credit for. This brings me to why I'm here.

The Issue: In my effort to build a dependency parser, I've come to really disagree with the Igbo language academia on a number of fronts. I'm very well familiar with the works of many Igbo linguists/language researchers, from earlier researchers like G. E. Igwe to contemporaries like O. N. Anyanwu. While I appreciate their excellent contribution (and great documentation), my efforts in building a dependency parser (based in part on their collective works) has left me feeling as though they failed to analyze the language for what it really is and based significant portions of their interpretation on semantic comparisons with English. So now, I am venturing on the edge of a few new hypothesis about some morphological and syntactic aspects of the language, and I want to try bouncing some ideas off others to see if I am heading in the right direction.

Now the Main Part: I want to administer a single test. Don't worry, it is not something that requires technical expertise. I would prefer you rely on your native competancy in the language. Essentially, use your intution as speakers to go through the various tests and provide your response in this thread. Thank you.

Question 1: In Igbo, we build relationships between nouns simply by stringing them one after the other. This creates a noun phrase. The order that we string them does not matter, so long as they satisfy an association in our native intuition. The following are examples of what I mean by such association.

"ime ulo" => "ulo ime"
"aka elu" => "elu aka"
"okpuru akwa" => "akwa okpuru"
"akwukwo iwu" => "iwu akwukwo"
"eze oha" => "oha eze"

I can go on, but I'm hoping the picture is clear, using your native intuition. The above transformations are all considered valid in our native intuition. The meaning in English is not what is important here, but rather the fact that the nouns can be switched and the association will still be considered valid in our native intuition, even though the meaning changes.

Now, I'm going to list a series of transformations similar to the above, please use your native intuition to mark them as either valid or invalid.

1. "ogbu agu" => "agu ogbu"
2. "mpa n'aka" => "n'aka mpa"
3. "mgbafu oso" => "oso mgbafu"
4. "ora ofe" => "ofe ora"
5. "mbido isi" => "isi mbido"
6. "ofe ogwugwu" => "ogwugwu ofe"

Question 2: If you did not mark any of the above as invalid, skip this question and go to Question 3. Igbo is an SVO language with strict word order. What that basically means is that there is always a subject (S) responsible for an action (V) that affects or is received by an object (O).

For example, "o gara ahia", where "o" is the subject, "gara" is the action being discussed and "ahia" is what is targeted or affected by that action.

This ties into our native intuition of "verbal force". Igbo is a strictly verb-based language and draws distinction between a "thing being referenced" (non-verbal force) vs an "action being discussed" (verbal force).

For each of the ones you marked as "invalid" above, do you believe any still contain "verbal force"? If so, which ones?

Question 3: On the topic of verbal force. Please label whether or not you believe any of the below have verbal force. Simply say "yes" for verbal force, or "no" for non-verbal force.

1. erimeri
2. mbadamba
3. apirapi
4. ebighebi
5. ahurahu
6. nchorincho

This will have to suffice for now. I do have a few other questions, but I'll wait for some meaningful engagement before I take it further.

After I went through your post,I got to know why you used the tag "seeking igbo language experts"...
Although I'm science inclined,I could hold my ground in history,but when it comes to literature and Igbo language,I am insignificant

UCHIHAmask, yes, I was concerned that a thorough reading might deter most. Truly, I don't think someone needs to be versed in literature and Igbo language research. Rather, what I am hoping for is that people will at least try to use their native intuition to answer (what I hope will be interpreted as) some simple questions. Perhaps I have myself to blame for thinking I was providing necessary context in each question, when all it probably ended up doing was scaring people off.

But seriously, give it a try and share your thoughts. Think of it as a survey. I just want to get the opinion of other Igbo speakers to see if there is a basis for the morphological hypotheses I am currently developing (of it is all just in my head).

ChinenyeN:
UCHIHAmask, yes, I was concerned that a thorough reading might deter most. Truly, I don't think someone needs to be versed in literature and Igbo language research. Rather, what I am hoping for is that people will at least try to use their native intuition to answer (what I hope will be interpreted as) some simple questions. Perhaps I have myself to blame for thinking I was providing necessary context in each question, when all it probably ended up doing was scaring people off.

But seriously, give it a try and share your thoughts. Think of it as a survey. I just want to get the opinion of other Igbo speakers to see if there is a basis for the morphological hypotheses I am currently developing (of it is all just in my head).

Yes! "Native intuition",that is exactly what is keeping me from accurately perceiving the context of your text,you see,only recently did I seriously take to aim my intent in perfecting Igbo language usage, originally I think myself to have achieved 70% of Igbo usage and that I was only lacking 30% foundation of not growing up in igboland,maybe I overestimate my level,maybe that foundation is something I need to do more about, coming to this culture section has made me realize that there is something more even after I attain 100% Igbo language usage.

You ain't really scaring off people (maybe people like me though), it's a level above me, I'm sure that there are 'molten magma' kind of Igbos who would find the interpretation of your hypothesis a Cake, my place now is as an audience....
Another thing! How did you tag me by just mentioning my username?

UCHIHAmask, we have the Nairaland developers to thank for this neat feature. They made it so that simply mentioning a handle triggers a notice. So tagging someone is as simple as that.

As for the rest of your post. While it is true that I am certainly seeking at least one person with a fundamental understanding of nominal compounding in linguistics as well as genitive vs non-genitive relationship between nominals, I would still also very much like to hear the thoughts and feelings of others besides myself that speak. Ultimately, I want to avoid the effects of keeping myself in an echo chamber and having my hypotheses sound good to me, but end up being nonsense when tested with others. That said, I understand your reservations. No wahala.

Well, since I've already managed to scare off aghalari unu, I guess I will just go ahead and expand a little more, in case anyone is curious about the context behind my request.

Basically, I develop software as a profession (programming, engineering, etc., whatever you want to call it). I have had the goal to develop software that will help boost the Ngwa lexicon by creating new words. With my understanding of Ngwa, linguistics research and software deveopment, I have managed to coin new lexicon that is currently in active use. However, it is not good enough. If I want to bring Ngwa back to the point of being a language of instruction, I'm going to need to create many more new words. I've discovered that the problem lies in our understanding of the morphosyntactic dependency structure of Ngwa (and Igbo at large). So now I am writing software to analyze and parse the dependence structure of Ngwa sentences. The simple goal is to bypass our modern orthographic approach at writing Igbo and take a morphosyntactic (grammarian's) approach at writing to see if I can uncover more of the thought process behind how different expressions were created in the past.

Now, I have come against a blocker. Attempting to define what constitues as a true noun + noun compound vs what constitutes as a genitive expression, vs what constitues as a verbal compound.

I am relatively certain that I have figured out the dependency structure for various types of verbal compounding. The real issue lies in the genitive vs non-genitive compounding of noun + noun multiword expressions. So I have come to nde we nnwe m la nde maa nnwe. Anybody that actively speaks Igbo and doesn't mind answer the questions identified in the original post would be pushing this project along a helpful path. In fact, I'm fairly certain that any principles I discover for coining new terminology can likely be applied to any other lect just as easily as with Ngwa. Who knows, maybe I'll help expand your own community's lexicon as well.

Okay, for those who may be interested in knowing more, here is why this issue of compounding is significant.

Every Igbo lect has lexicon that seems to have an opaque etymology. Take the word "nwatakiri" or "nwakatiri" for instance. This is actually an excellent example of why this issue of compounding is significant.

"nwatakiri" is obviously treated as a noun in its own right, but for all speakers, it is clear that "nwatakiri" is the agglutination of at least two nouns, "nwa" + "ntakiri". If we dive deeper, as speakers, we even know in our intuition that "ntakiri" is the agglutination of "nta" + whatever "kiri" is supposed to have been. "kiri" is the opaque part of this word.

"nwatakiri" represents one of the many examples of a "true compound" according to linguists like Emenanjo, and they make a good point. "nwa" + "nta" + "kiri", when rendered as "nwatakiri", indicates a rather flat association between "nwa" + "nta" + "kiri". None of the hallmarks of a genitive association exists.

According to Emenanjo, the morphosyntactic process that allowed for the transition of the genitive "nwa" + "nta" + "kiri" to the compound "nwatakiri" is no longer productive among the majority of Igbo lects. In other words, most Igbo lects have completely lost the ability to create new words with compounding. Practically every "noun" + "noun" multiword expression in modern Igbo is in the genitive case. Emenanjo does admit though that there are lects in which this morphosyntactic process is still somewhat productive.

If we can expose and revive this morphosyntactic process across the board, we can in fact unlock the ability to productively create a myriad of new lexicon. Plain and simple.

So I wanted to test a hypothesis about how we perceive certain types of "noun" + "noun" association as Igbo speakers to see if I can uncover a pattern that proves (or disproves) my hypothesis on how to revive the morphosyntactic process.

If the “nwatakiri” example is not clear enough, let us look at another example, “nwa” + “nne”.

Most Igbo will try to explain the etymology of “nwa” + “nne” using the genitive case and claim that it means “child of (the) mother”. While it may be true that the etymology might have come from the genitive case in antiquity, by the time we arrive at modern Igbo, it had since transformed to a compound. Now, rather than it meaning “child of (the) mother”, “nwanne” is actually better translated as “childmother”, the genitive “of” sense gone.

It may be difficult to see initially, but in our intuition as speakers, we know it is true. So how do we prove it? Easy, parse the dependency relationship. Let’s look at the dependency relationship of “nwanne m” vs “nwanne m”. Yes, I know. Those are the same, you want to say, but they aren’t. One is the compound and the other is the genitive case.

In the genitive case, the “m” modifies “nne”, giving us “child of my mother”. See how the “my” is associated with “mother” solely. In the compound case, “m” modifies the entire “nwanne” expression, giving us “my childmother” (as a term of kinship or endearment). One need not actually be “child of my mother” to be “nwanne m”.

To further cement this, let us examine the final piece of compounding, phonological changes (even shifts). For example, “nwa nta” -> “nwa ta”.

Even though we write “nwa” + “nne” as “nwanne”, the pronunciation in a number of lects is actually a bit different.

In places like Ngwa and Mbaise, it’s as though it is pronounced as “nwane”, you can also hear “nwune” in parts of Mbaise. In places like Ikwerre, Echie and Asa, we see “nwene”. In Ekpeye, we get “nwana”. In some communities in the Isu complex, we see “nwune”. In Echie, Ngwa and Mbaise it can be taken a step further to “nweneta” (Echie) or “nwaneta” (Ngwa and Mbaise) for a junior.

We can easily see the difference between the genitive case and the compound when comparing across all lects.

Ngwa - nwa nne (genitive) - nwane (compound)
Mbaise - nwa nne (genitive) - nwane/nwune (compound)
Ikwerre - nwo nne (genitive) - nwene (compound)
Echie - nwa nne (genitive) - nwene (compound)
Asa - nwa nne (genitive) - nwene (compound)
Ekpeye - unwo ina (genitive) - nwana (compound)
Isu - nwa nne (genitive) - nwune (compound)

The genitive case still exists in all the above lects, proving that the phonology change shows that the genitive relationship is gone in the compound. It then follows that we can use the weakening of the phonetic boundary between words as evidence of true compounds. Rather than the individual nouns being treated as individual nouns in their own right, they are treated collectively as a single noun, opening up them up internally to phonetic shifts between themselves.

I wouldn’t be surprised if we saw other such examples among other lects. It’s probably also the case that other lects also pronounce “nwanne” as “nwane” or maybe even used to pronounce it a different way, but modern Igbo orthography has reinforced writing “nwanne” and forced the compound out of the lexicon.

The fact that the genitive “nwanne” is reinforced in modern Igbo further signifies that modern Igbo has (in large part) lost the capacity for true compounds.

Alright, let me show the dependency graph to perhaps should help ginger our intuitions. I'll still use the expression "nwannē m". I'll provide two dependency trees. One showing the genitive relationship and the other showing the compound relationship.

- nwa
    |- nnē
          |- m

The above is the genetive relationship. In other words, "child of mother of me". In other words, "my mother's child". Here we see that "m" directly modifies "nnē" which in turn directly modifies "nwa". If you were to ask any modern Igbo person how to explain what "nwannē m" means, this above is undoubtedly the explanation that they will give. Why? Because modern Igbo can seemingly only conceptualize the genitive case.

But, when we actually make the statement "nwannē m" in speech, we never actually mean "my mother's child". Instead, we mean something else. We mean "my nwannē" and not "nwa, my nnē".

- nwa
|    |- nnē
|
|- m

The above shows the compound relationship between "nwa" and "nnē" and we see that "m" does not modify "nnē" but rather modifies the entire compound "nwannē". This transition to a compound opens us up to phonological changes in "nwannē", and we finally arrive at the below.

- nwa
|    |- nē
|
|- m

As well as all its various lectal variants.

This suggests that proto-Igbo made very productive use of this morphosyntactic process. So many lexical items that we inherited from proto-Igbo a the result of this very productive morphosyntactic process which is now lost among many Igbo-speaking communities.

Just to further show the example with "nwatàkịrị".

- nwa
    |- ntàkịrị
        |- ihe

vs

- nwa
|    |- tàkịrị
|
|- ihe

These two show two different relationships with "ihe" modifying "ntàkịrị" in the first inatance and "ihe" modifing "nwatàkịrị" in the second. The difference in meaning is clearer with this example. To say "nwa ntàkịrị ihe" is to say "a little something". To say "nwatàkịrị ihe" (child of something).

This is why true compounding is so important.

Maybe an even better example of what I mean by compounding is the expression "madụ̀". It has the halmarks of a compound. We can identify its separate parts, and most especially, we have one part with an opaque etymology. This probably also makes for a better example of a compound because the genitive expression no longer exists. Here's what I mean. If we examine different lectal variants of "madụ̀" we see we can quickly see that the "dụ̀" is the word for "life". However, the etymology of "ma" remains opaque.

madụ̀ == whatever "ma" is supposed to be + "ndụ̀"

According to the work of linguists Ohiri-Aniche, Blench and Williamson (the linguists that did proto-Igbo reconstruction), the "ma" in "madụ̀" might have originally be a "mgba", based on their comparative study across surviving Igbo lects. If we take their linguistic research seriously, we might then say that the original genitive form of "madụ̀" might have been along the lines of "mgba ndụ̀" (mgba of life, a live mgba, a living mgba). We can try to deduce any number of interpretations for "mgba", but that is beyond the scope of the point being made here. Perhaps if anyone wants to try to dig into that later, we can.

Anyhow, the point is that "mgba ndụ̀" went from having a genitive relationship "mgba of life" to a compound relationship "mgba life" and that soon opened it up to internal changes in its phonology. The "gb" weakening to "b" in some lects and perhaps weakening further to an "m" in some, while disappearing altogether in others. The "n" in "ndụ̀" being clipped off.

Through this weakening of the phonetic boundary between words, we get the following (potential) branching.

"mgba ndụ̀" => "mba dụ̀"

"mba dụ̀" => "ba dụ̀" => "va dụ̀"

"mba dụ̀" => "mma dụ̀" => "ma dụ̀"

"mba dụ̀" => "ma dụ̀"

If we were to take a grammarian's approach toward writing Igbo, we should actually be separating the "ma" and "dụ̀" and writing it as "ma dụ̀", because we know that it is two identifiable nominals compounded into a "multiword expression" without the genitive case. This final part (without the genitive case) is the most important part of compounding, I believe. This is one of the major features that differentiates modern Igbo from proto Igbo.

The underlying relationship looks like this:

- ma mgba NOUN
  |- dụ̀ ndụ̀ NOUN

Comparing "ma dụ̀" to "mgba ndụ̀" side-by-side, we should be able to see with our native intuition that the genitive case that exists in "mgba ndụ̀" is completely missing in "ma dụ̀".

Yeah, I think this is a good example since the the genitive case no longer exists in modern Igbo, having been completely displaced by the compound it seems. So to resurrect the genitive case now would show how contrasting it is from the compound that has survived it.

Honestly, think about it. Let us say that we had lost 90% of our lexicon and are trying to revive the various modern Igbo lects, and that for some reason the word for "person" is one of the lexical items that was lost. Let's also say that modern Igbo had to coin a new expression for "person" and ended up with something like "mụọ̄ ndụ̀".

In "mụọ̄ ndụ̀" we already see the genitive relationship. "mụọ̄ ndụ̀" give us the idea of "mụọ̄ ǹkè ndụ̀". In fact, for those who don't understand the genitive case in linguistic studies, you can think of it as adding "ǹkè" between nouns. In fact, yes, this should definitely help show the contrast between the genitive and the compound.

If the multiword expression uses "ǹkè" or has an implicit "ǹkè", it is genitive. Removing the "ǹkè" and removing even the implicit notion of "ǹkè" give us the compound.

So think about "mụọ̄ ǹkè ndụ̀" and "mụọ̄ ndụ̀" (with zero notion of ǹkè). That is the genitive vs the compound, respectively. If we followed the path of our ancestors, the next step would be to weaken the phonetic boundary between "mụọ̄" and "ndụ̀", and that would give us any number of lexical items like "mụdụ̀", "mọdụ̀", and because the boundary is weak, we can even further assimilate the expression and maybe get "mọdọ̀" in some lects.

It is that simple.

If modern Igbo did not lose the capacity for true compounds, we could potentially see many such examples new lexical items in modern day using this morphosyntactic/morphophonolocial process. It is unlikely that the weakening did not follow some sort of rules, even if they were followed loosely. Identifying these rules and putting them back into practice can help revive the intuition for compounding. This is ultimately the goal of my parser.