Is there a difference between plurality in semantics and in morphology?

Question

With regard to morphology a common example of a lexeme is [dog, dogs] where dogs is the plural inflexion of the lemma dog modified by the -s suffix, marking plurality.

Although I can accept that dog and dogs are, morphologically speaking, the same word i.e. the same lexeme, it still bothers me because each has a different meaning.

Semantically speaking, can we say dog and dogs are the same words? How does plurality relates to dog semantically:

1. does dogs means multiple instances of a class dog?
2. or does it mean the multiplicity of a same instance?

To illustrate the different possibilities here above:

1. "Suddenly several dogs, not one, were surrounding me."

2. "Suddenly several Benjamins, not one, were surrounding me."

In the second example, I used a proper name to emphasize the individuality of the instance, which is multiplied.

In the former case, if dog is understood as a class, then dogs could refer to multiple instances of that class as well as to multiple sub-classes of that same class. By the same logic, dog (singular) could mean a class or an instance of that class.

"The dog will always be man's best companion." (class)
"The dog entered my yard this morning." (instance)

Is the semantic distinction between class and instance here a shortcoming of the English language? Could a hypothetical language have distinct words for dog-class, dog-class-single-instance, dog-class-multiple-instances, etc.? Another example which may suggest the distinction in English would be fish:

"One fish",
"Two fish",
"Tuna's a large fish",
"I breed several fishes"

What does plurality mean semantically and how does it reconcile with plurality in morphology?

• Could one say forest is the semantical plural of tree?
• How about words which have a plural that is morphologically entirely different from their singular (e.g. person, people)?
• How about words which take a different meaning with a plural inflexion (e.g. F. vacance [vacancy], vacances [holidays])?

Please provide examples in your answer.

Answer 1

I can maybe answer some of the questions raised above. Although I probably won't provide the kinds of examples you are looking for. There are at least two ways that people go about representing the semantics of plurality, sets or sums.

Link 1983 introduces the notion of a sum to the ontology. So in addition to having individual objects in the ontology, eg. b which stands for the dog Bingo, and a, which stands for the dog Angel, you have sum, or plural individuals, which correspond, eg., to a+b, the sum of Bruno and Angel. So the singular word 'dog' would denote the set of singular dog individuals, {a, b}, and the plural form 'dogs' would denote then set of singular AND sum/plural dog individuals, {a, b, a+b}.(Well, some people would say it denotes the set of just plural/sum individuals, {a+b}, but this is a controversial issue.) Usually you would have more individuals in your model, eg. a singular noun could denote {a, b, c, d}, and the pluralized noun denotes {a, b, c, d, a+b, a+c, a+d, b+c, b+d, c+d, a+b+c, ..., a+b+c+d}. When you have what you are calling class-uses of the noun (Carlson called this a kind-use), eg. "The bear is a carnivore", the noun is often formalized as denoting the maximal individual in the set, i.e., the individual that is the sum of all of the elements in the set, a+b+c+d.

As for 'forest', I think most approaches would not say that it is the semantic plural of 'tree' (which would be 'trees') - but that there are singular individuals in our ontology which correspond to things that we would call a forest. So 'forest' would have a denotation like 'dog', but a corresponds to The Black Forest, b corresponds to The Haunted Forest, etc. There might be operations (like Landman's Group operator) that map plural individuals in the denotation of 'trees' to the singular individuals in the denotation of 'forest', but these are still represented as distinct individuals semantically.

As for (morphologically) irregular plurals, Link introduces this * notation. So the plural 'dogs' would be semantically represented as *dog, where this *-operator is what maps from a singular's denotation to a plural/sum denotation. This * need not be confined to the presence of a specific plural morpheme, however, so even forms which have suppletive plural forms would be represented with the *, i.e., the denotation of 'children' would be *child. So I think morphological plurality, in English at least, differs from semantic plurality. (And of course we have pluralia tantum like 'pants' and 'scissors', which really suggest against having no difference between morphological and semantic plurality, as noted by Knitter above).

The set approach, I am not as familiar with, but think it is one where the singular noun 'dog' again denotes the set of singular individual dogs {a, b}, but there is no new type of individual introduced to the ontology. Instead, the plural form 'dogs' denotes the set of all the possible sets of dogs (the power set), {{a}, {b}, {a, b}}. (again, I guess some people would argue that it is this set minus the singleton set members). (Schwarzschild 1996 is what I ought to have consulted previous to answering this, but...)

Under either approach to plurality, a singular and plural form have distinct denotations, so I don't think they would be considered the same word semantically (assuming that the notion of a semantic word would correspond to a denotation, although I don't know if that is what's usually done.)

Answer 2

I'm going to try to answer the spirit of the title question, although I have no formal training semantics and thus cannot answer some of the more specific questions you pose in the body of the answer (I'm more interested in morphology).

Simply put, there is often a disconnect between semantics and morphology, and it is something of a quixotic endeavor to try to fully reconcile them.

However you define plurality, there are cases in which the usual markers to denote it do not correspond to a concept of plural, as you've pointed out. For example, "scissors" is morphologically plural (this is evidenced by verb agreement: the scissors are ...) but semantically refers only to one object (a case of plurale tantum). This is probably also what is happening with vacances: the use of vacance in the plural took on a specialized meaning which then became fossilized as a plurale tantum. These cases are problematic even if one accepts that there is a morphological "zero" plural in English which accounts for pairs of ambiguous plurality like deer-deer, fish-fish. Add in collective nouns, which are morphologically singular though they may be considered "plurally" semantically or even in other parts of the morphosyntax (e.g. "My family is/are happy to see you") and what seems to be a (relatively) straightforward semantic notion of plurality becomes quite muddled in the morphology indeed! (Also, there is the phenomena of plural attraction, in which a nearby plural noun can trigger plural agreement with the verb even though the full subject is singular, for example before proof-reading I had written: "...there are cases in which the usual markers to denote it does...".)

These facts all point to the idea that plurality as a morphosyntactic relation is distinct from semantic plurality. There is usually a relationship between a semantically plural referent and plural verbal agreement or nominal morphology, but not always: morphological plurality can combine with semantic singularity (pluralia tantum) and morphological singularity of the noun can combine with plural semantics and even plural agreement (e.g. the case of "family"). How speakers mentally reconcile this, I do not know.

It would be nice if every discrete piece of morphology had only one semantic or syntactic meaning. In fact, analyzing words and then breaking them down into parts based on shared semantic contribution is, roughly speaking, how morphemes are analyzed. But the mapping is hardly one to one. As handy as semantics can be in analyzing morphemes, it has been a problem knowing how to reconcile cases in which the semantics and morphology don't match up. A great case are "Priscianic" or parasitic formations, in which one member of a paradigm seems to be formed not on the lexical root of the paradigm but instead on the stem of another member of the paradigm. In Latin, the future active participle in the vast majority of cases is based off of a stem identical to the stem of the supine and past (usually passive) participle. Obviously, future and active ≠ past and passive participle, and so how do you analyze the semantic status of the stem?

Answer 3

The main question you're asking has been addressed directly in a recent paper on the relationship between morphological and semantic markedness.