Stephen van Vlack
Sookmyung Women`s University
Graduate School of English
Introduction to Linguistics
Fall 2004
Week 6: Questions for Singleton, Chapter 3
1. What are the building blocks of meaning in language?
According to Singleton, even though words seem to enjoy some sort of special status in language they do not seem to be the basic units in language. One main reason for this is that it is very hard to tie down exactly what a word is. Even in languages that have what are easily recognized as word units it is quite hard to always find the boundaries between parts. The basic idea is that words are more often than not complex entities composed of many different parts. Based on this it would seem that it is morphemes and not words which are the main building blocks of meaning.
In order to understand this better let`s take look at some different types of morphemes and morphological operations. The simplest distinction between morphemes is between so-called free and bound morphemes. Free morphemes are forms that are able to stand on their own, while bound morphemes cannot. In English the words porcupine or armadillo are said to be free morphemes because they can stand on their own and they are also able to have inflections like plural markers tacked onto them. These tacked on elements are bound morphemes.
(1) John wanted sure if there was just one armadillo or many different armadillos.
(2) The famous recipe sadly called for the inclusion of not just one whole porcupine but several partial porcupines.
Based on such examples it seems we can make a clear distinction between free morphemes which can stand alone and bound morphemes, like -s, which cannot. The distinction however might not be nearly as simple for all languages as it might at first glance appear for English. In class this week we also introduced the idea of different categories for words, dividing them into lexical phrases and functional phrases. It would be wonderful if the bound/free distinction correlated strongly with the functional/lexical distinction we discussed in class but unfortunately it does not. Bound morphemes are both lexical and functional in that some of them have a clear lexical meaning while others only have a functional purpose. Likewise some free morphemes belong to lexical categories (NP, VP, AP, PP) while others belong to functional categories (CP, IP, DP). This is shown below.
(3) Free Morphemes
wander (VP) `Try not to let your thoughts wander as you read this.`
shoe (NP) `John threw his shoe at me.`
fatal(AP) `The armadillo was involved in a fatal accident.`
above (PP) `The new drama coach comes from heaven above.`
while (CP) `She fell down while she was trying to walk and chew gum at the same time.`
would(IP) `I was wondering if you would mind not smiling at me.`
that (DP) `The elephant wanted to sit on that person.`
(4) Bound Morphemes
(VP) re- `We are hoping that she will reform her behavior.`
(NP) -tion `We are currently going through a kind of revolution in linguistics.`
(AP) -ic `The place is so romantic I just can`t leave.`
(tense) -ed `Bill wanted to eat peas.`
(agreement) -s `June answers to the name Snookums.`
(aspect) -ing `Stop making such stupid examples.`
These different types of morphemes are involved principally in three major types of so-called productive operations in English derivational affixation, inflectional affixation, and compounding. Affixation involves attaching bound morphemes either to the left most (prefixation) or right most (suffixation) boundary of a free morpheme. Compounding involves combining two free morphemes, again from either boundary.
(5) Affixation = [?util[Vize], [?util[Nity], [?util[Viza[Ntion], [Vre[Voccur[Nence]
(6) Compounding = [Ablack[Nbird], [Nseat[Awarm[Ner]
Now, any description of morphological operations in English is going to be reliant on the clear distinction between free and bound morphemes, but this distinction is far from clear. In several of the samples used above, porcupine, armadillo, utility, it is quite unclear if all these morphemes are indeed free morphemes. Porcupine, for its part is made up of 2 parts, `porcu` and `pine`. Now, we recognize the word `pine`, but the word `porcu` is strange to us. Even pine here has a different meaning in the one that we are used using in relation to trees.`porcu`, seems familiar in being related to the word pork but the form is different and the meaning is different. It is therefore hard to claimed strongly that this word, porcupine, is the result of something morphological operation in English. Clearly it is not. It may have been in Latin but in English is not. The same can be said for armadillo, which is of Spanish origin.`arma`we recognize as having something to do with weapons or armor, but the second part `(d)illo` is a Spanish morpheme which makes things small (this is called a diminutive). Thus it is clear for English that this word as well is not part of any kind of morphological operation in English, although it certainly was in Spanish. All these examples might seem far-fetched and certainly English is filled with thousands of them. Linguists would argue that words like porcupine and armadillo are not examples because they're not the result of productive processes. Certainly this is true but if we look more at the example utilize above we find a similar problem. In all the examples with the base `util` given in (6) above we are left wondering about the base form. While the space seems to be employed in many different productive morphological processes, it is not at all clear that the base really is a free morpheme in English. It doesn't exist anywhere except in morphologically complex forms, yet it is a base.
In Korean we find a similar situation with the base `hak`. This form seems to behave like a free morpheme in the morphological operation that it is involved in, yet it never appears alone. Certainly phenomenon like this in both Korean and English is very common and principally for the same reason. Both languages have borrowed extensively from neighboring prestige languages. Keep this problem in mind as we will discuss it further when we get down to question four and the whole idea of how morphemes might be stored in the mental lexicon.
2. What are allomorphs and how do they relate to the lexicon?
Simply put, allomorphs are slightly variant forms of the same morpheme. The basic idea is that there is one idealized form of each morpheme which is stored for language use. At the same time, however, the actual form of the morpheme may or may not vary depending on the linguistic environment in which defines itself. A simple example would be the morpheme /-in/ which has the variant forms [-un] and [-im]. Basically [-un] and [-im] are allomorphs of /-in/. /-in/ is the usual form but in some places [-un] or [-im] will occur, again depending on environmental conditions. This phenomenon is called allomorphy and it is very common particularly in languages with a large amount of affixation.
(7) Allomorphy of /-in/
[in[Ndivis[Aible]
[im[Nposs[Aible]
[un[Vdo]
Linguists claim that allomorphy is conditioned; that is is created by certain lingusitic environments. Based on this belief is, therefore, possible (in fact it must be possible) to create rules from which allomorphs are derived from the actual morpheme. In the above example it is quite easy create a phonological rule deriving [-im] from /-in/ based on a former being followed by a word starting with a bilabial consonant. The case of [-un] is less clear and therefore much more problematic. It could be argued that [-un] is a different morpheme which just happens to share a similar shape to /-in/. We make his proposal based on the observation that /-un/ seems to have a different function than /-in/. This is how linguist go about trying to figure out how morphemes might actually work.
3. What is the difference between lexical and inflectional morphology?
Linguists, and here we are primarily talking about Matthews, often make a clear distinction between what is termed lexical and inflectional morphology. The basic idea is that lexical morphology is somehow idiosyncratic and therefore must be stored somewhere in the mental lexicon. This means that all complex word forms must be stored in the mental lexicon. Its runs in strong contrast to the views of some linguists, most notably Di Sciullo and Williams who claimed morphological processes, even derivational morphological processes such as we find in examples (5) and (7) above are constrained by elements of universal grammar and therefore productive and need not be contained in the lexicon. Obviously there are conflicting views on everything including this. Singleton makes his point clear by pointing out that even in derivational processes which may seem productive there are changes on base forms caused by morphological processes which are not entirely productive. Some bound morphemes have effects on vases while others do not and the effects vary. Singleton, again following this simple distinction made by Matthews (1972; 1974), argues that all derivational processes are somehow encoded in the lexicon, although he doesn't discount the possibility that there might be some rule-based conditions which apply.
Inflectional morphology, on the other hand, has generally been claimed by linguists to be rule-governed; that is, their use occurs based on rules which are used to produce new forms through these morphological operation, hence the use of the word productive. Since inflectional morphology is claimed to be productive there is then no need for inflectional forms to be housed in the mental lexicon. A simple example would be the way in which English uses inflection to mark number on nouns. It is claimed that there is a single plural morpheme /-z/ with two allomorphs [-s] and [-iz], which are derived from the morpheme /-z/ through some simple phonological rules. All this is productive and rule governed. Therefore, in this argument, there is no need to have separate lexical entries for plural forms in English, except of course for plural forms which do not follow this rule. In such cases of idiosyncratic plural formation, such as foot - feet, then the plural forms must be specified in the lexicon. All inflection is claimed under this view to be essentially rule-governed and therefore inflected forms (due to their predictability) need not be stored in the lexicon like other forms derived through `lexical` morphological processes.
4. How are inflections proposed to be stored in the lexicon?
A careful study, however, of more highly inflected languages reveals that inflections are not necessarily very predictable, as they are in English.
Here are some noun forms present in the English of approximately 700-1100 AD.
Masculine Noun
Singular `the king`
Nominative se cyning
Accusative thone cyning
Genitive thaes cyninges
Dative taem, thy cyninge
Plural
Nominative tha cyningas
Accusative tha cyningas
Genitive thara cyninga
Dative thaem cyningum
`the day`
daeg
daeg
daeges
daege
dagas
dagas
daga
dagum
`the place`
stede
stede
stedes
stede
stede
stede
steda
stedum
Neuter Noun `the ship`
Singular
Nominative thaet scip
Accusative thaet scip
Genitive thaes scipes
Dative thaem, thy scipe
Plural
Nominative tha scipu
Accusative tha scipu
Genitive thara scipa
Dative thaem scipum
`the punishment`
wite
wite
wites
wite
witu
witu
wita
witum
Singular
Nominative seo glof
Accusative tha glofe
Genitive thaere glofe
Dative thaere glofe
Plural
Nominative tha glofa
Accusative tha glofa
Genitive thara glofa
Dative theaem glofum
`the deed`
daed
daed
deade
deade
deade
daede
daeda
daedum
The basic idea here is that all forms not only need to be in the lexicon, as Singleton proposes in this chapter, but forms must also be connected somehow. If complex forms must be in the lexicon, what about less complex forms, the building blocks? Is there a lexical entry for bases like `util` or `hak` which are productive but do not exist on their own? Well, the answer that Singleton gives us is a resounding YES. Everything must be in there. This question of size will be further developed next week.