The subject notion and Functional Ways of Structuring Language

Term Paper (Advanced seminar) 2007 16 Pages

English Language and Literature Studies - Linguistics


Table of Contents

1. Introduction

2. The notion ‘subject’
2.1 The category ‘grammatical subject’
2.2 Prototypicality of the grammatical subject-category

3. Functional approaches and the notion ‘subject’
3.1 Topic and comment
3.2 Given and new information
3.3 Figure, ground and perspective

4. Conclusion


1. Introduction

In this term paper I will inve stigate the structure ofEnglish sentences with the subject notion as a starting po int. It pres ents a clas sical notion to analyse clau ses and s entences but how exactly can a subject be defined? For this purpose, I will show that the notion is not detailed enough and suggest a distinction in to grammatical, logical and psychological subject. This proves useful to analyse sentences which at first glance do not appear to have any subje ct ata ll. Ina next s tep I will focus on f eatures of the gram matical subject according to the Cambridge Grammar of the English language (2005). The discussion will prove that the properties given for grammatical subjects do not constitute a fixed fra me which sharply distinguishes between elem ents eligible to be subjects and others that are not. Instead I will argue that the sub ject category is best analysed as a prototype ca tegory and that its features have prototypical character.

The second section is concerned with di fferent ways of accounting for particular structures oflanguage. If various syntactic functions can appear at the beginning of sentences then why does a speaker choose a part icular construction in stead of another? I will argu e that this question is closely re lated to analys es of claus es, senten ces and utterances g oing beyon dam eres ubject vs. p redicate dic hotomy. I will s tart w ith a discussion of the thematic structure of sentences and clauses and introduce the distinction of topic and comment. The second step com plements the them atic structure oflanguage with the inf ormation structure, in which c onstituents can be labelle d ‘given’ and ‘new’. This analysis also considers the intra- and extra-linguistic context of clauses and sentences and can th ereby acco unt for a fair sh are of speaker-choices b etween d iffering constructions. Since there are still so me cases that cannot be explained b y looking at the information structure, I will then present the notion of perspective as very helpful. Taking together these different levels of analysis one is enabled to account for a large qu antity of possible constructions in the English language.

2. The notion ‘subject’

The category ‘subject’ is in itself not quite cl ear and before I start the discussion I will try to give a definition. Ei senberg (2006: 280-283) discusse s the difference between the terms semantic orlogical subject, psychologi cal subject and grammatical subject. While the gramm atical subject is quite a comm on notion, the other two are not. According to Wagner (2007) the logical subject can be defi ned according to its sem antic role. Usually the logical subject would occupy th e role of an agent but if there is no ag ent in a sentence it is likely to occupy other sem antic role s following a hierarchy proposed by Bates & MacWhinney (1982: 200): agent > experiencer > dative > instrument >patient > location

The term psychological subject is a more functional term and I will discuss in the next chapter which terms are appropriate to describe it. For a distinction between grammatical and logical subject, compare the following sentences in which I underline the grammatical subject:

(1) Rooney scored a goal.
(2) A goal was scored by Rooney.
(3) It was raining during the match.
(4) (You) Go home!

In (1) we clearly have a grammatical subject “R ooney”, which is identical with the logical subject, i.e. the initia tor of the action. T his m eans that the grammatical and the logical subject are com bined in the sam e syntactic element. However, gramm atical and logical subject do not always ha ve to coincide. In (2) the gr ammatical subject “a goal” is not identical with the agent “Rooney”. In this case we call “Rooney” the logical subject to express that the e lement is the initiator of th e action. The distin ction between th e two aspects of the term subject p roves particularly useful in sentences like (3) and (4). In (3) we can identify a grammatical subject “it” but there is no logical sub ject in the sentence. This is why the pronoun in this case m ight be called non-referential it. In the im perative sentence (4) it is the other way round. W hile no gramm atical subject can be identified there is certainly a logical subj ect that the im perative refers to, e.g. “you” as indicated in brackets. T he exact referent will only be comprehensible from the context of the sentence.[1]

2.1 The category ‘grammatical subject’

The questio n arises ho w exactly the gramm atical sub ject can be identified in a sentence or clause. Huddleston (2005: 236-239) states that in most cases the subject will be an agent and express the pres entational status of topic. Then he gives a list of further “distinctive grammatical properties of the [s c. grammatical] subject” considering clause internal f eatures in can onical constructions. I will present these p roperties inas lightly adapted version here:

a) Category: prototypically a NP

b) Position: usually before the predicator
c) Case: nominative if morphologically realised
d) Agreement with the form of the verb
e) Subject-auxiliary inversion in closed interrogatives and some other constructions
f) Subject-auxiliary-inversion in open inte rrogatives if the inte rrogative element is not the subject.
g) Tags agree with the subject
h) Coordination: VP-coordination is possible
i) Obligatoriness: subjects are oblig atory in can onical clauses; subjectless clauses only in non-canonical constructions
j) Uniqueness: only one subject per clause Huddleston claims that this catalogue is va lid for canonical clause s. He then applies the properties to several non-ca nonical constructions with th e result that subjects in non- canonical constructions also fulfil the given properties. There isjust the m inor exception that some of the tests cannot usefully be applied. As for the rem aining ones the subjects behave in the predicted manner (cf. Huddleston 2005: 239-244).

2.2 Prototypicality of the grammatical subject-category

I argue, however, that at leas t the properties a), b), c), d) and i) are controversial and that considering them as prototypical features in the sense that e .g. Taylor (1998) and esp. Bates & MacWhinney (1982) suggest is m ore appropriate. They suggest that prototypical members of a category fulfil all the criteria and that less prototypical members might only fulfil som e of the m. Consequently, ther eisa degree of“goodness of me mbership” according to howmany features overlap with the prototype (Bates & MacWhinney 1982: 211). The margins of the category are not clear-cut but all members are supposed to bear a “family resemblance” with the p rototype or ha ve a “central ten dency” towards it. Since Huddleston claims that his subject properties account for all possible subjects in English, I will argue with examples containing both canonical and non-canonical sentences.

Before starting with his list of subject prope rties, Huddleston m entions that in m ost cases the subject will be an agent and express the topic of a sentence. As for the sem antic role of the gramm atical sub ject, I woul d refer back to the h ierarchy by Bates & MacWhinney (1982: 200), I mentioned earlier. They convincingly argue that if a sentence contains an agent, th is is lik ely to be th e grammatical s ubject; but if not, then other semantic roles are ready to step in, like for instance in (5) where an ins trument occupies the gramm atical subject function. This shows that the feature “agent” has prototypical character:

(5) The ball hit the goal keeper.

The same goes for the topicality of the subject, which is a prototypical feature as well, but not a distinct one: Not al l subjects are top ics of sen tences! I will provide further information about the term topic in 3.1 and explain it in more detail.

Starting with his list of subject properti es, Huddleston writes that the gramm atical subject is “prototypically” a NP and that subordinate clauses and other elem ents occur in subject position - if only under some se mantic and syntactic restrictions.[2] The f ollowing sentences provide only two possible exam ples for further elem ents in the grammatical subject position, namely a PP and an AdvP:

(6) In the goal is a badplace to hide.

(7) Quickly is howyou have to dribblepast the defender.

The word order criterion according to b) can be relativised by a closer look at the list: considering e) and f), it shows that a word order is possible that deviates from the prototypical version which has the subject in clause-initi al position. Furthermore, the non- canonical construction of subject postposing constitutes a second grammatical clause type in which the subject does not p recede the predicator. While Huddleston gives a tho rough discussion of the phenom enon of the two constr uctions, he does not explic itly admit that his prop erty b) revea ls proto typical inst ead of uncond itional char acter. Still, thes e constructions provide evidence for the prototypi cal function of criterion b), even if not many deviations from the prototype are a llowed. Sentences (8) and (9) with subject- auxiliary-inversion and subject postposing illustrate this fact:

(8) Did the football player kick the ball?

(9) Very disappointing were theplayers’ excuses after the match.

InHuddleston’sargum entc)concerning thecaseofgrammati calsubjectsitis important to note that he only considers fi nite sentences. Firstly, one m ight mention different forms of infinitive clauses in which the grammatical subject is in the ob jective case. Only allowing the nominative case for grammatical subjects would make an analysis of the infinitive clause in (10) difficult. Secondly, if someone objected to the foregoing by saying that clauses cannot stand alone and ha ve to be treated differently, one could mention the incredulity response (cf. Taylor 1998: 186f.). It is ap parent that this construction needs a context, but it is gram matically possible and the pronoun reveals that the subject of the respo nse stands in the objective case. Compare the following sentences in which I indicate the infinitive-clause with square-brackets:

(10) It was not very clever [for him to insult his coach].

(11) Him score a goal!?!

Just like for Huddleston’s pr operty b) there are two argum ents for the prototypicality of c) concerning the ag reement between grammatical subject and verb. Firstly, one could again refer to the incredulity response (11) or infinitive clauses in which no agreement is established. Secondly, in some cases of informal speech there is no agreement between the verb and the gramm atical subject (12). Tomy knowledge this phenom enon occurs frequently with the auxiliary ‘do’, less f requently with ‘b e’ and rar ely with ve rbs. This supports the claim that the boundaries of the subject category are not completely clear-cut:

(12) %I’m sure Rooney don’t mind!

Thelastelem entfromHuddleston’slist thatIwillm entionisi)concerningthe obligatoriness of gram matical subjects. W hen he s ays th at ev ery sentence contains a subject, if not present then implied, Huddleston might be referring to a logical subject in the case of the implied subject. Be it or not, one can presume that in prototypical sentences there is a gramm atical subject. Howeve r, there are non-canonical sentences like imperatives (13) or non-finite clauses (14) where no gramm atical subject is present. I insert a possible logical subject in brackets and indicate the infinitive-clause in (14) with square-brackets:

(13) (You)Gohome!

(14) Itisa sign ofpoorperformance [(for afootballplayer) to lose the ball].

As a conclusion of this section one can sa y that features for a distinct category, containing all possible grammatical subjects, are difficult to establish. Most of the subject properties given by Huddleston are not absolute ly excluding as I ha ve shown, only g), h) andj) from his list seem fairly clear-cut . This observation suggests that the subject category is more appro priately described in term s ofa prot otypical category. Particularly Huddleston’s properties a), the kind of elements occurring in subject position, and b), their position in a claus e, ar e open to d eviation f rom the prototypica l ca se. His prope rties together with the two aspects ‘agent’ and ‘topicality’, which he only m entions in passing, form an apt description of the prototypical grammatical subject.[3]

3. Functional approaches and the notion ‘subject’

When Huddleston applies his subject properties to non-canonical structures, he proves thatinpreposing, subjectpostposingand subject-auxiliary-inversionthegram matical subject does not chang e com pared to the canon ical version, but th at it does so in the passive, as well as in extrapos ition and existen tial constructions. However, with a m ere sentence-/ clause-internal analysis, Huddl eston cannot explain the choice between different constructions. In som e cases th e non-canonical sentences can have the same proposition as their canonical counterparts. Compare the examples:

(15) That Beckham doesn’t score goals is aproblem.

(16) Itisaproblem that Beckham doesn’t score goals.

The question arises why a speaker would choose one construction or another. This is what the f ollowing sections are con cerned with. I will outline dif ferent ways of thematic and information s tructure in sen tences a nd discuss their connect ion with the notion subject.[4] Here, the term ‘psychological subject’ comes into play again.

3.1 Topic and comment

A first way to inves tigate why sp eakers pr efer one construction of a sentence to another is to look at the them atic struct ure of sentences. This approach does not distinguish between the subject an da predicator of a clau se but betw een its topic and comment.[5] The topic of a sentence can be define d as “what is being talked about” w hile the comm ent is “the point being m ade about that topic” (Bates & MacW hinney 1982: 197). It is important to note that this distinct ion is usually a sentence-internal one and can in m ost cases be determ ined without consid ering the context of a clause or sentence.

English as opposed to othe r languages like Japanese a nd Mandarin has no m orphological coding to mark the topic (Li & Thompson 1976: 466), which is why diffe rent criteria for the identification of the topic have to be esta blished. Firstly, the topic is always the first meaningful constituent in a sentence (Bloor & Bloor 2004: 72). Accordingly, interjections or vocatives would not fill the topic-slot. As the first element it is in the focus of attention and the whole comment-sequence refers to it. Secondly, the topic can have different forms in the surface structu re of a clause. In othe r words, it can coincide with the gramm atical subject of a sentence (17) - and in unmarked constructions it usually does - but it can also be realised by the grammatical functions adjunct (18), complement (19) or predicator (20) (Bloor & Bloor 2004: 76). I have underlined the topic in the example sentences:

(17) David Beckham easily scored this goal.

(18) Easily David Beckham scored this goal.

(19) This goal David Beckham scored easily.

(20) Score this goal, David Beckham!

Thirdly, the example sentences also show another aspect. The topic does not have to be an argument of the verb or in fact, it does not have to be connected with the verb at all. As a consequence, there does not have to be any agreement between the verb of a sentence and its topic. However, if there is such an agreement of topic and verb then this is simply a matter of one particular instance. Fourthly, Bates & MacW hinney (1982: 201) distinguish different layers of the topic-comment relationship within one utterance. There can be more than one such relationship inas entence and the comment of the first clause can function as the topic of the next one, whose comment can again be the topic of the following clause etc.

As a conclusion one can say that the gramm atical subject does not have to be identical with the to pic but it m ight be in unm arked constructions.


[1] Durscheid (2007: 186f.) points to the problem that in sentences like (3) the element ,,it“ has to be classified assubject(gra mmatically)and non-subject(lo gically)atth esa me time.Sheconsidersthisa mere terminological problem and can stick to the idea of such a distinction. I would like to add that this ‘problem’ also occurs in sentence (2). However, it is the purpose of this approach to open the opportunity of classifying an element [± grammatical subject] and [± logical subject], which means that the effect is intended.

[2] Huddleston does not explain the term “prototypical” and his way of applying the category features suggests clear-cut boundaries forthe category ‘subject’ instead of prototypicality.

[3] Cf. Newmeyer (2000). He argues against the possible prototypicality of any lexical or syntactical category. Even though his argumentation has a point he is very selective in the prototype theories he discusses.

[4] The t erms topic/ co mment, figure/ ground, theme/ rheme, given/ new are used by different authors with varying meanings. I follow the terminology as suggested by Durscheid (2007).

[5] Again there is some terminological confusion. Bloor & Bloor (2004) refer to what I call topic and comment as theme (ortopical theme) and rheme. However, my use of t he terms is supported by many other authors: e.g. cf. Bates & McWhinney (1982), Durscheid (2007), Li & Thompson (1976), Ward (2005). I will not go into the details of Bloor & Bloor’s distinction between different forms of topic (2004: 77-79).


ISBN (eBook)
ISBN (Book)
File size
498 KB
Catalog Number
Institution / College
University of Heidelberg – Institut für Anglistik
Functional Ways Structuring Language




Title: The subject notion and Functional Ways of Structuring Language