Imperatives or Word—Class Conversion？

【Abstract】Some subjectless sentences introduced by a bare verb caught my attention. English language has low tolerance for pro-drop phenomenon. Hence， I assume these subjectless sentences are not declarative clauses. The paper analyzes this linguistic phenomenon from the perspective of Governing and Binding Theory and borrows some concepts from Functional Grammar. The conclusion is：（1） they can be regarded as lets-elided imperatives with advising function； （2） the introductive verbs show features of conjunction and preposition， which implies inflectionless word-class conversion may occur.

【Key words】pro-drop； imperative； conversion

1. Introduction：the puzzle

1a Suppose Im beautiful. Id always be haunted by the feeling that you had been taking a chance on just that， and that kind of love would disgust me. （Text A， Unit 6， New College English Book 1）

1b Consider the time Charlene Jendry， a conservationist at the Columbus Zoo， learned that a female gorilla named Colo was handling suspicious object. （Unit 7， Text A， New College English Book 1）

1c Assume the year 2001 is 365 days and that for some reason most days on a certain ratio measure is around July 15th. （Multiple Regression and Its Discontents， Education， 2012）

In the above sentences， suppose， consider and assume are conventionally used as verbs， which means the subjects are missing （or seemingly missing） in the sentences. In Chomskys view， English is not a pro-drop language （Haegeman 1991）. How to explain the above phenomenon？ In this paper， an analysis to explain these subjectless sentences are conducted based on pro-drop， imperatives and conversion analysis.

2. Literature review

2.1 Previous studies on pro-drop phenomena in English

Unlike Italian and Spanish， which have rich inflection for person and number， and unlike Chinese and Japanese， with almost no inflection， English is something in between. According to Cook and Newson （2000）， languages like English with poor inflection should be categorized as non-pro-drop languages. However， English does allow null-subject in finite clauses under certain circumstances， which have been observed and studies by linguists.

2.1.1 pro-drop in imperatives

The null-subject pro in imperatives was first studied. Sweet （1960：111） wrote， “as the imperatives can be used only in addressing someone， the subject of an imperative sentence must always be in the second person.” So far， its believed that the pro in imperatives has the feature of the second-person pronoun.

2.1.2 pro-drop in Early Modern English

Radford （1997） suggests that Early Modern English （EME） could be a pro-drop language. The evidence in support of his claim comes from the following instances：

2a Hast any more of this？ （Trinculo， The Tempest， II， ii）

（Modern English：Have you any more of this？）

2b Canst not rule her？ （Lenontes， Winters Tale， II， ii）

（Modern English：Cant you rule her？）

2c Wilt come？ （Stephano， The Tempest， III， ii）

（Modern English：Will you come？）

Since inflection in EME is richer than todays modern English， the finite verbs have stronger agreement with their subjects in person and number. The –st inflection in （2a） （hast = present tense second-person singular of have） and （2b） （canst = present tense second-person singular of cant） indicate that the subject is“thou”， a second-person singular subject. Similarly， -t inflection in （2c） also reveals that the subject is in second-person singular form.

2.1.3 pro-drop in Informal Writing and Colloquial Styles

Haegeman （1990） gave detailed account of covert subjects in diaries； Hyams （1993） and Rizzi （1986） investigated the subject-missing sentences in early child English. Subjectless clauses are also found in telegrams （Barton 1998）， slogans， commercials （Ozaki 2010） and colloquial speeches （Thrasher 1977）.

However， the previous studies can not give a satisfactory answer to the question：how does pro-drop take place in textbooks， which should be written in standardized modern English？

2.2 Imperatives

Quirk （1972）， from the perspective of addressee， divided imperatives into three major types：first person imperative， second person imperative and third person imperative. It can be showed in the following table：

1ST PERSON 2ND PERSON 3RD PERSON

without subject __ Open the door. __

with subject without let __ You open the door. Someone open the door.

with let Let me open the door. Lets open the door. __ Let someone/him/them open the door.

The above table indicates that a subjectless imperative does not take a first-person and third-person pronoun as its covert subject， which is also stated in Sweets book （1960）， but when let is placed at the beginning of the sentences， the imperative does not take a second-person pronoun as its subject. An imperative does not take third-person pronouns he/they as its subject， but it can take the indefinite pronouns someone/somebody as its subject.

In Quicks explanation （1972）：when we apply imperatives， we intend to command， require， permit or prohibit the addressee（s） to act. Only the second-person pronoun you can be used in this case （someone or somebody can refer to any of the addressee in the contextual implication）. In the case of let-headed imperatives， the covert subject is also the second-person pronoun you.

It is worth noticing that INFL（TENSE/AGR） disappears in imperatives：

3a Someone open the door.

3b *Somebody opens the door.

（Note：* indicates ungrammaticality）

With regard to the functions of imperatives， Sun （2005） categorizes imperatives into five types：order， request， advice， criticism and answer.

There are several contextual variables affecting the functions of imperatives：

1. Speaker – the initiator of the imperatives. The speaker can be individual（s）， social organizations， government authorities and so on.

2. Addressee – the performer of the proposed action of the imperative. The addressee can be certain individual（s） and general audience.

3. Motivation – whose needs the speaker issues an imperative for.

4. Status – the symmetrical or asymmetrical relationship between the interlocutors. The relationship is symmetrical if the interlocutors are equal （+/-） and asymmetry if they are unequal. In the case of asymmetry， the speaker is either of higher status than the addressee （S+； A-） or of lower status than the addressee （S-； A+）.

5. Beneficiary – the person who will benefit the fulfillment of the proposed action.

6. Option – the freedom of the addressee to comply or refuse to perform the proposed action.

7. Consequence – the loss the addressee will suffer or punishment he will face if the addressee refuses to comply.

The relation between the contextual variables and the functional types is shown as follow：

（Note：“S” refers to“the speaker”；“A” refers to“the addressee”；“+” means“possessing the feature”；“-“means“not possessing the feature” and“+/-” means“either possessing or not possessing the feature”.）

2.3 Word-Class Conversion

Conversion is a way in which an English word can shift from one word class （part of speech） to another without changing its form. There are three ways of changing a words class， namely affixation， backformation and conversion （Li 2016）. Affixation is attaching a morpheme （affix） to a word stem to form a new word or word form， while backformation is the opposite process in which the actual or“supposed” affix（es） is/are removed from a word to create a new lexeme. Conversion is a word shifting from one word class to another without changing its morphological form.

4a All this， needless to say， had been culled second-hand from radio report.

4b The students wanted more say in the government of the university.

The word say shifts from a verb in （4a） to a noun in （4b） without adding something to or removing anything from its form. Unlike the other two methods， conversion does not require any change in form.

However， conversion normally occurs in adjectives/adverbials/propositions to nouns， nouns/adjectives to verbs and verbs to nouns （Li 2016）. The conversion of verb to conjunction is hardly seen in current available literature.

3. The analysis of the null-subject sentences

3.1 Are they simply non-imperative pro-drop clauses？

Our intuition tells us that sentences （1a-c） are more likely to be imperatives than non-imperatives， but we still begin our analysis from the least possibility.

First， according to theta-criterion， each argument must have a thematic role and each thematic role must be assigned to an argument. The verb suppose in （1a） is a two-place predicate， and it should have two thematic roles to assign. However， we can only find one argument in the sentence：the complementiser phrase （CP） Im beautiful. There must be a covert argument in the sentence. Second， Extended Projection Principle （EPP） requires a subject in a sentence. Then what is this covert argument， or the null subject in the sentence？

If we determine that there is a covert argument in （1a）， the sentence can be rewritten as follow：

5 [IP1 [e] suppose [CP[IP2 I am beautiful]]].

（Note：[e] symbolizes empty category.）

There are four main types of empty categories：NP-trace， Wh-trace， PRO， and pro.

NP-movement can occur in a clause （e.g. passivization） or move an NP subject from a lower clause to a higher clause （NP-raising）， but never an NP subject from a higher clause to a lower clause. Thus， [e] in （5） is not the trace of NP I：

*[[e]t suppose [It am beautiful]].

There is no wh-constituent in （5）， so the empty category [e] cant be wh-trace.

PRO only appears in non-finite clauses and the sentence is a finite， so [e] is not a PRO. The only option left is pro （Similar analyses on sentences （1b-c） draw the same conclusion）. Could it be pro？

6 [e] Seems she grew up with four younger brothers and sisters. （Maxwell， Murder at Beechwood， 2015）

INFL（TENSE/AGR） in EME pro-drop sentences still finds its trace in colloquial English spoken today. We can infer from sentence （6） that [e] is the dummy subject it， which shows tense and number agreement with the tense inflection –s. Thus， we can accordingly assume that the verbs in pro-drop clauses retain such tense and agreement inflections.

7a *Supposes/*Supposed Im beautiful.

7b *Considers/*Considered the time Charlene Jendry， a conservationist at the Columbus Zoo， learned that a female gorilla named Colo was handling suspicious object.

7c *Assumes/*Assumed the year 2001 is 365 days and that for some reason most days on a certain ratio measure is around July 15th.

Pro-drop clauses need to retain inflections in the verbs for the necessary null-subject inference. The ungrammaticality of V-es/V-ed +CP in sentences indicates that these sentences are not pro-drop declarative clauses， because inflected verbs are not allowed in them.

3.2 Are they subjectless imperatives？

They are not declarative clauses. Could they be imperatives？ Then what persons do the null subjects take？

If lets or let me is inserted at the beginning of the sentences， we get sentences：

8a Lets/*Let me suppose Im beautiful.

8b Lets/*Let me consider the time Charlene Jendry ...

8c Lets/*Let me assume the year 2001 is 365 days and that for some reason most days on a certain ratio measure is around July 15th.

Similarly， if we put you or someone before the sentences：

9a *You/*someone suppose Im beautiful.

9b *You/*someone consider the time Charlene Jendry ...

9c *You/*someone assume the year 2001 is 365 days and that for some reason most days on a certain ratio measure is around July 15th.

How come lets is more idiomatic than let me， you and someone in the sentences？ We can analyze the contextual variables of these imperative markers to see which types they can fall into：

Request is an imperative the speaker initiates out of his own needs； the speaker is always of lower status and the addressee has the power to refuse compliance （Searle and Vanderveken 1985：199）. We can see that the sentence *Let me consider Im beautiful （=Please allow me to consider Im beautiful） is most likely to be a request. The virtual action performer is the speaker himself/herself； the addressee is just the performer of allow. A request must benefit the speaker and answer his/her needs. However， the speaker does not benefit from the proposed action in the sentence. With simple logical reasoning， we know that the addressee should not be excluded from the action performers of consider.

8a/2 Lets suppose Im beautiful.

Advice （advising or suggesting） is an imperative in which the speaker gives his opinion to the addressee about what to behave and expects the addressee to perform the proposed action. The addressee has the freedom to reject the proposed action whose fulfillment benefits the addressee， and sometimes benefits a third party （ibid）.

Since lets includes the speaker and the addressee， the speakers intention of saying it is to suggest the addressee make an action， in our example， an assumption which benefits the addressee in understanding the speakers assumption.

9a *You/*someone suppose Im beautiful.

In imperative of order， the speaker is in a position of power over the addressee， and the addressee has little freedom to refuse to comply or he must face the consequence （ibid）. The speaker tends to use words like someone and you as sentence subjects， in a rude and unfriendly manner. As we can see， order is not contextually proper here. The speaker is not rude and is not enforcing compliance on the addressee to carry out his/her action.

Based on the above analysis， “Suppose Im beautiful” is most likely to be an advice.

Some linguists just divide imperatives into two distinct groups. Bach and Harnish （quoted from Davies 1986：35） draw a distinction， basing on the requirement of compliance， between requestives and requirements. A requirement is desire of the speaker for something to be done； a requestive is an imperative where the speaker need not desire compliance at all， but intends his utterance itself to motivate the addressee to obey. Haverkate（1986） distinguishes imperatives basing on whether the speaker is appealing to his authority over the addressee. He divides them into two groups：impositive， in which the speaker possesses authority over the addressee； non-impostitive， in which the speaker does not possess authority over the addressee.

Imperatives Suns Classification Bach and Harnishs dichotomy Haverkates dichotomy

1）*Let me suppose Im beautiful.

2）Let me open the door. Request requestives

requirements non-impositive

impositive

1）Lets suppose Im beautiful.

2）Lets open the door. Advice

1）*Someone suppose Im beautiful.

2）Someone open the door. Order

1）*You suppose Im beautiful.

2）You open the door. Order

As we can see from the table， Suppose Im beautiful should not be at the extreme ends of the arrows. It cant be a request （at the end of requestives/non-impositive） nor an order （at the end of requirements/impositive）. It should be something in between， which testifies its advising feature discussed before.

To sum up， if sentences （1a-c） are imperatives， they are probably lets-elided imperatives with advising function.

Should let take a covert subject preceding it？ According to some linguists （Quirk et al 1985）， let has evolved into a particle， or pure imperative marker， losing all of its verbal features. In the case of imperatives， let is seen as an adjunct， theta-criterion does not apply to it.

4.3 How about conversion？

Some dictionaries， like CALD3 give suppose a new word class – conjunction， sharing same meaning and same word class of supposing：

10 Suppose/Supposing we miss the train – what will we do then？ （CALD3：2269）

This example among many indicates that conversion occur in the traditional verb suppose， it can substitute its conjunctional derivative supposing.

However， verb-to-conjunction conversion is rarely seen and seldom draws much attention. Besides， such conversion is not found in the entries of consider and assume – even though considering and assuming both leave an entry in dictionaries as conjunctions.

The hypothesis of regarding suppose， consider， assume as conjunctions can powerfully explain the V[-tense][-AGR] + CP structure， but it has a flaw – it cant explain consider + NP structure like sentence （1c）：

1b *[ CP1 [CONJ Consider] [NP the time [CP2 Charlene Jendry， a conservationist at the Columbus Zoo， [V learned] [CP3that a female gorilla named Colo was handling suspicious object]]]].

If consider in （1b） were a conjunction， it would leave CP1 without a predicative verb. But it does not mean verb-to-conjunction-shift hypothesis is a weak explanation to this linguistic phenomenon in that language has been and is changing of all time. In a transition， an emerging feature will cause the dying out of an existing feature. Suppose shows the feature of supposing in V+CP structure， because suppose can take a clause as its object but not a noun phrase； while V+NP/CP feature （meaning consider can take both noun phrase and clause as it object） prohibits consider to show the conjunctional feature in V+NP structure. In this structure， V+NP is an imperative.

The gradual replacement of suppose for supposing may find some support in Leechs prediction （1995）：English is changing from an inflected language （synthetic language） to an isolating language （analytic language）. The affixation and inflection may progressively fade out. Such“deflexional” changes have taken place in the phenomena of some adjectives replacing their –ly adverbials， like“slow” being used as its adverbial“slowly”， verbs replacing their noun forms， such as“assist” for“assistance” and“invite” for“invitation” in colloquial English.

5. Conclusion

Sentences （1a-c） can be regarded as advising imperatives in which the imperative marker lets is elided. They can also be considered as sentences introduced by a verb converted into a conjunction （except 1b， in which the noun phrase prohibits the conversion）.

The subjectless V[-INFL]+CP structure in discussion seems to occur with mental-state verbs， such as consider， suppose， assume and imagine. The covert subjects are the experiencers of the mental act. Be an experiencers reference clear or ambiguous， it does not affect the meaning of CP， which is the theme of the sentence. Most of these verbs have -ing inflected derivatives as conjunctions or prepositions. Its worth further investigation， which is not done in this paper due to my limitation.

References：

[1]Barton，E.L.1998.The Grammar of Telegraphic Structures.Journal of English Linguistic.26：37-67.

[2]Cook，V.J.& M.Newson.2000.Comskys Universal Grammar： An Introduction.Cambridge，Mass：Blackwell Press.

[3]Davies，E.1986.The English Imperative.New Hampshire：Croom Helm Ltd.

[4]Haegeman，L.1990.Understood Subjects in English Diaries.Multilingual 9：157-199.

[5]Haegeman，L.1991.Introduction to Government and Binding Theory.Oxford：Blackwell Press.

[6]Haverkate，H.1986.A Model for Analyzing Politeness Strategies in Verbal Interation.Paper presented at the XIth World Congress of Sociology，New Dehli.

[7]Hyams，N.& K.Wexler.1993.On the Grammatical Basis of Null Subjects in Child Language.Linguistic Inquiry.24：421-459.

[8]Leech，G.1995.English Grammar - Past，Present and Future.Language Teaching and Research.2：1-6.

[9]Li，Qian.2016.A Cognitive Study of Conversion of Enlgish Adjectives.Hunan Normal University.

[10]Ozaki，S.2010.Subjectless Sentences in English.Nagoya Bunri University 10：35-46.

[11]Quirk，R.Greenhaum，S.，Leech，G.and Svartvik，J.1972.A Grammar of Contemporary English.London：Longman.

[12]Radford，A.1997.Syntax：A Minimalist Introduction.Cambridge：Cambridge University Press.

[13]Rizzi，L.1986.Null Objects in Italian and the Theory of pro.Linguistic Inquiry.17：501-557.

[14]Searle，J.R.and Vanderveken，D.1985.Foundations od Illocutionary Logic.Cambridge：Combridge University Press.

[15]Sun，Hui Zeng.2005.Funtional Analysis of Imperatives.Hebei Teachers University.13-17.

[16]Sweet，H.1960.A New English Grammar，Logical and Historical.Oxford：The Clarendon Press.

[17]Thrasher，R.1977.One Way to Say More by Saying Less：A Study of So-Called Subjectless Sentences.Tokyo：The Eihosha.

作者簡介：蔡健智（1981-），男，漢族，廣東人，講師，碩士，研究方向：語言學，二語習得。