When to split off morphemes

[Toby A]

This may not be the right forum to ask this question since it's not entirely to do with FLEx, but I can;t find a better forum of people qualified to discuss this, so if someone knows of a good technical linguistics forum please direct me there.

When I'm analysing texts and glossing morphemes there are cases like adpositions and compass directions that clearly have the locative affix on them, but I think the stem without the affix wouldn't make any sense on its own. Should I split off the locative affix in the word for "north" for example, and if so how would I gloss the stem?

Maybe in these cases I should let the whole word go into the lexicon in one piece, which makes me think about other situations where I can divide the word into sensible morphemes, but perhaps shouldn't. For example the word for "where" is the word for "what" with the locative affix, and the word for "here" is the near demonstrative with the locative affix. I've been splitting these word into morphemes in the text and words analysis tab, but now I realise that means the lexicon and resulting dictionary wont have "where" and "here" as their own entries, and maybe they should.

Any thoughts on this?

Toby

[Richard Gravina]

For me the best way to approach it is to decide whether you are treating the processes as lexical derivation or as part of the grammar. If these are not highly productive processes, you may be better off handling them within the lexicon. E.g. you would have an entry for ‘where’ in the lexicon, and make it a Complex Form, derived from ‘what’.

 

One that I find difficult is how to treat prepositional pronouns such as from-me, with-you, in-it etc. It may just boil down to whether you want them to appear in the dictionary or not.

 

Richard Gravina

--
You are subscribed to the publicly accessible group "FLEx list".
Only members can post but anyone can view messages on the website.
---
You received this message because you are subscribed to the Google Groups "FLEx list" group.
To unsubscribe from this group and stop receiving emails from it, send an email to flex-list+...@googlegroups.com.
To post to this group, send email to flex...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/flex-list/7b19ae6f-fba4-4c8f-856d-6462f4fe7bb8%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Ron Moe]

To understand the difference between inflection and derivation, go to Andy Black's paper under Helps-Resources-Introduction to Parsing and look at section 2.1.5. He gives some tests to help you decide if an affix is inflectional or derivational. He also tells you how to handle them with the parser.

Generally inflectional affixes are added to your larger grammatical categories (noun, verb, adjective). But they can also go on smaller categories too. But with small categories like adpositions it is highly unlikely that an affix is inflectional.

The distribution of derivational affixes is limited and irregular. On the other hand inflectional affixes can occur on almost any member of a grammatical category and most stems can be fully inflected. When you find exceptions, they usually belong to a class of words and there is a semantic reason for the exception. For instance "scissors" does not occur in the singular. It belongs to a small class of things like "shears", "pants", and "binoculars" that have two parts. These words only occur in the singular in compounds like "pant leg" and "binocular vision".

The normal rule is to treat inflection as non-lexical. That means that we don't include inflected forms in the lexicon. But we do treat derivation as lexical and include derived forms in the lexicon. Your locative affix sounds very much like a derivational affix. So you should put all the forms into the lexicon.

Andy's paper doesn't go into much detail about the semantic differences between inflection and derivation. The basic rule is that inflection adds grammatical meaning and derivation adds lexical meaning. Here is some additional material that I've written that may help you sort out the difference between inflection and derivation.

Ron Moe

----------------

With inflection the meaning of the root stays the same. So adding -s to "chair" to form "chairs" means that we are now talking about more than one chair. But the meaning of "chair" stays the same. Likewise adding -ed to "listen" to form "listened" merely puts the listening into the past. It does not change the meaning of "listen". In both words the meaning of the root is the same whether it is singular or plural, present or past.

With inflection the meaning of the affix also stays the same. In both "chairs" and "listened" the affixes -s and -ed have their regular meanings 'plural' and 'past'. Like most things in language, you can find exceptions. (If you add -s to a mass noun, it means 'a type of X', e.g. fruit:fruits.) But generally the meaning of the affix remains stable. The meaning of the inflected word is the sum of the two parts.

But with derivation the meaning of root plus affix is often unpredictable. In English the suffix -er usually means 'a person (or thing) who does X'. Pairs such as help:helper and teach:teacher follow the pattern. But we can also find pairs like dream:dreamer and catch:catcher in which the meaning has shifted. Sometimes the derivative has a different meaning or an additional meaning. In the pair rule:ruler "rule" means 'to govern' and "ruler" means 'a person who governs'. But ruler can also mean 'a flat stick with marks along the edge indicating units of length; used to draw straight lines and to measure the length of things'. The word "rule" does not have an equivalent meaning of 'to measure the length of something'. This kind of shift in meaning is common with derivational affixes, but extremely rare with inflectional affixes. 

The reason for this is that the root and the derivative are different words and can develop in different ways. But a stem and its inflected forms are the same word. When the meaning of a word changes over time, all the inflected forms change together. But a derivative is only loosely tied to its root. The meaning of a derivative can change so much that we find it hard to see the connection. A stretcher doesn't stretch. Instead cloth is stretched over a frame. We can see the connection between an officer in the army and an office, but the connection isn't straightforward. Other kinds of changes can occur as well. We can recognize a similar pattern in message:messenger and passage:passenger, but the phonological pattern is limited to these two words and the semantic relationship isn't the same. A root and derivative can diverge to the point where the two are no longer associated in the mind. For instance the word "ladder" is inherited from the Proto-Indo-European root *klei-. The inherited verb form is "lean". "Tinker" is probably from "tinkle".

When we come to the smaller grammatical classes we have to be careful. They can be inflected, but usually are not. Even when we see regular patterns, they should usually be treated as derivational unless they follow a similar pattern to the larger grammatical classes. In some languages pronouns and demonstratives are inflected like nouns or adjectives. Regular patterns that are rare should be entered into the lexicon as derivational. For instance -ward can occur on numerous location words to derive a locative adverb with the basic meaning 'in the direction of': fore:forward, on:onward, backward, rearward, upward, downward, inward, outward, northward, southward, eastward, westward, windward, leeward, skyward. But you can't suffix -ward to any location word: *behindward, *sternward, *sideward, *awayward. Neither is the meaning entirely predictable. "Afterward" has a temporal meaning and "wayward" means 'disobedient'. "Toward" has a shortened pronunciation. So in spite of its productivity -ward is clearly derivational.

--

[Mike Maxwell]

On 3/9/2015 7:01 PM, Ron Moe wrote:
> with small categories like adpositions it is highly unlikely that an affix is
> inflectional.

There's a discussion of person marking on adpositions here:
http://wals.info/chapter/48
It does note that it's not always clear whether a person marker on an adposition
should be considered to be an affix or a clitic (or perhaps an independent
pronoun). There are some other examples here, where the claim is that some
head-marking languages allow inflection on adpositions:
http://en.wikipedia.org/wiki/Inflection
Welsh has been claimed to be another language with inflected prepositions:
http://en.wiktionary.org/wiki/Category:Welsh_prepositions
There's a better description than the wiktionary one on a pdf at
http://privatewww.essex.ac.uk/; unfortunately, Google won't give me the real
URL, just a https://www.google.com address, which is worthless. Ah, wait, Bing
is better behaved:
privatewww.essex.ac.uk/~louisa/esrcproj/WelshPrepositions.pdf
Here the inflections can be followed by an explicit pronoun, so it's clear that
the inflections really are inflectional affixes, not clitics.
--
Mike Maxwell
max...@umiacs.umd.edu
"My definition of an interesting universe is
one that has the capacity to study itself."
--Stephen Eastmond

[Joyce Wood]

This is almost exactly my question. But I have a specifically FLEx-related aspect of the question.

I know that the suffix in question (-en) in my FLEx is derivational, that it affects the lexical semantics of the verb it attaches to. I have determined that I want to represent it in the Lexicon (via complex forms / components) but not in the morphemes analysis in Texts & Words, that it would be better to not separate the -en off, better to create a new lexical entry for the derived verb.

However, when I go to make the change, I notice that I have 100+ instances of this derivational suffix in my interlinearized texts. I really appreciate the ability to run a concordance search on a specific morpheme/entry/etc in my texts, and I worry about the loss of this feature, if I implement this decision to represent the entire/derived word instead of the verb+suffix. Is that true? If I do not analyze the morpheme break because the suffix is derivational, will there be a way to run a concordance search on verbs which contain this suffix? Or will there just be the list in the lexical entry of -en of the entries which I have manually indicated contain this suffix?

(Also, I'm assuming that I will have to manually change the analysis for those 100+ examples in my texts which are currently analyzed with this suffix. Please let me know if there is a quicker way to do this.)

Joyce