Carwash and rent-a-car a typological investigation.

Carwash (1) is not found among the headwords of OED. Rent-a-car is referred to under rent (V1), the date of the first quotation of the word being given as 1924. According to WCD, carwash is an "area or structure equipped with facilities for washing automobiles". Rent-a-car is explained as having come from (imper.) rent a car. Can we not say, too, that it comes from a suggestion or offering, i.e. "we will rent a car"? Both [WID.sub.3] and RHD contain car wash but treat the expression as two orthographic words. [WND.sub.3] gives the item in two variant forms, carwash and car wash. The meaning given in RHD is "a place or structure, as at a service station, having special equipment for washing automobiles". It is similar in [WID.sub.3]. But the headword rent-a-car is found in none of them. In [COD.sub.8] neither of these is found as a headword. Only in the item of [rent.sub.1] rent-a- is mentioned, with two exemplifying words, one of which is rent-a-van. [COD.sub.9] includes these two as headwords, car wash (n) a nd rent-a-. Under the latter item we find rent-a-van again.

In Lightfoot (1979: 160) carwash is given with windowcleaner as a possible example of the so-called OV compounds in terms of linguistic typology. In contrast, pickpocket and other similar words are regarded as belonging to the VO type. Many of these latter examples are followed in parentheses with the dates of the first quotations in OED. Pickpocket has the date 1591. The outline of Lightfoot's assertion is as follows.

Windowcleaner, for example, belongs to the OV type of compounds, which are harmonious with the SOV order of the sentence. For the past eight hundred years, at least, this SOV order has not been basic, but compounds of the type have been produced. Lightfoot argues as follows: if the element order of the compound reflects the basic order of the former age, the VO compounds of the Elizabethan age, exemplified as contrasts to the OV ones, should represent another SVO stage older than the Elizabethan period. He continues to say that "there is no evidence for such a claim and much evidence that there never was an earlier SVO stage" (Lightfoot 1979: 160). Is he criticizing typology with this argument?

The establishments signified by carwash and rent-a-car are popular and indispensable in our everyday life. The two compounds are frequently used. The former nominal compound, in my conjecture, (2) first came into being as car + washing, which was then comprehended as carwash + ing through metanalysis. (3) At the next stage, we have a verb compound to carwash through backformation. Then a homonymous noun was derived from the verb with a zero suffix, or produced by conversion with the meaning 'the place for car-washing'. On the other hand, rent-a-car is a nominal compound cut off from the so-called Verb Phrase (VP), and sometimes used as a modifier of another noun, for example, a rent-a-car company.

The process of sentence word order change in English may be roughly traced as follows:

(Proto-Germanic) (OE) (ME and ModE)

SOV [right arrow] {SVO SOV} [right arrow] SVO

Suppose the compound nouns, such as windowcleaner and carwash (OV), were still normal and dominant. The Modern English element order in the sentence is in the third stage of the diagram above. Concerning this order opposition between sentence (SVO) and compound (OV), the matter seems to be settled by referring to Givon's famous saying "Yesterday's syntax is today's morphology" (Givon 1971). It is often quoted, and we see it again in Lightfoot (1979: 160). In fact, Lehmann (1972: 245) reconstructed the normal order of Proto-Germanic, based on the construction of the compounds in Old English, Gothic, Old Norse, Old Saxon and Old High German. In Old English we find folc-agend (= one who owns people, i.e. chief), which is an OV compound. But we have some others which are formed on "today's syntax": we have a noun breakfast derived from a verb phrase to break (one's) fast.

Bauer considers the pattern Noun + Verb not so productive in noun compounding (1983: 205). But Lightfoot lists many compounds of this type, asserting as follows: "recent forms like car-wash indicate that this is still a productive pattern ...". He regards breakdown, etc. as "older VO forms" (1979: 160). In a very detailed way Akmajian discusses synthetic compounds, (4) that is, "those two-word English compounds in which the second word is the verbal" (1990: 42-45). As for the compound pattern Verb + Noun, he only enumerates three examples in Table 2.1 (1990: 24). In considering the type Verb + Noun, even Bauer admits that the "type used to be common ... If this pattern is still productive, it is only marginally so" (1979: 205).

It is difficult "to distinguish between compounds and syntactic phrases" Lehmann 1993: 255). In other words, the element order of compounds has a close connection with syntactic order of sentences through phrases. Phrases are small pieces cut off from sentences. At any rate, typology seeks as much harmony as possible at any level, and between the levels, of syntactic constructions of any particular language.

It might be possible to reason as follows. We cannot deny that the element order of compounds may change, following the new basic pattern of the sentence, but we must admit it will take some time for the change to be completed. Before the completion of this process, we may be sometimes at the mid-stage where we have two patterns of compound making side by side. But in the element order of compounds, if a new type is blocked by an older one after some period of prevalence, we must seek a satisfactory explanation, historical or typological, for the process. Otherwise typology will cease to be persuasive. Croft (1990) has some references to compounds, but does not seem to have great concern in their element order. We would like to examine the whole problem of word order from another angle.

Vennemann (1972/1974: 79) (5) gave us a table of various syntactic elements distributed into two classes, i.e. operator and operand. (A simplified revision of this table was shown in Comrie (1979: 92). Canale (1976: 40) showed it in a form faithful to the original.)

         A                   B
         (operator)          (operand)

  I (a)  object              verb
    (b)  adverbial           verb
    (c)  main verb           auxiliary
    (d)  main verb           modal
    (e)  main verb           intensional verb
 II (a)  adjective           noun
    (b)  relative clause     noun
    (c)  number marker       noun
    (d)  genitive            noun
    (e)  numeral             noun
    (f)  determiner          noun
III (a)  adjective stem      comparison marker
    (b)  standard            comparative adjective
    (c)  adverbial           adjective
 IV (a)  noun pharase        relation marker (adposition,
                             i.e. postposition or preposition)
  V (a)  indirect object     direct object
    (b)  temporal adverbial  directional adverbial

In Comrie (1979) Group V is deleted when showing Vennemann's table. Vennemann's second version of this table is quite similar to Comrie's revision.

A                              B

Object                         Verb
Adverb                         Verb
Main Verb                      Modal
Noun Modifier                  Noun
(Adjective, Relative Clause,
Adverbial attribute, Genitive
attribute)
Standard of comparison         Comparative adjective
Noun phrase                    Adposition
                               (Preposition, Postposition)
                                      Vennemann (1974: 345-346)

We have only two ways of ordering elements from column A and column B, AB and BA, in other words, operator + operand and operand + operator. The matters we should not overlook in the table are: (a) we cannot find the item of subject here, and (b) the first line in these tables, i.e. OV or VO, is regarded as the most fundamental, influencing other combinations. The 'object' is a cover word, including any modifier of the finite verb except the subject. The abbreviation of the object is O or X. According to Vennemann, in a particular language, if the order of the finite verb and its modifier is OV, that is AB, all other element orders would be generally AB. If the finite verb and its modifier stand in the reversed order (VO, that is, BA), all other constructions are expected to be in BA order. Vennemann explains this phenomenon as the analogy to the order of verb and object, as is evident in the title of his (1972/1974) paper.

We must not forget Lehmann in typological studies. Lehmann, too, considers the combination of verb and object as the nucleus of word order in the sentence. He regards the connection between subject and verb as on a different dimension than that which operator and operand belong to. In this respect Vennemann had the same opinion as Lehmann, or we may say, he followed Lehmann in former papers. The development of this idea is briefly observed in my former dissertation in Japanese (Fujiwara 1995: 185-187).

Vennemann gives his third version of the table of operator and operand (Vennemann 1982: 35). (6) We shall point out a few items different from those in his former versions of the table.

        A                       B
        (operator)              (operand)

I       ADVERBIAL               VERB
   (a)  subject, object         verb
   (b)  complement clause       verb
   (c)  adverb                  verb
   (d)  infinitive, participle  auxiliary verb
   (e)  infinitive              modal verb
II      ADNOMINAL               NOUN
   (a)  adjective               noun
   ...
 V      SENTENCE RADICAL        SUBJUNCTION
VI      SENTENCE RADICAL        MOOD MARKER

As is referred to above, Vennemann had asserted before that the construction of subject and predicate (verb) should belong to quite a different dimension than that of operator and operand. But it seems rather odd that the placement of elements in linear order should be on two different dimensions in one basic sentence. It would be proper and natural for the subject to have a position in this table of operator and operand. The verb is the nucleus of the sentence and the subject should fall under the heading of the operator, as does the object. This new measure of his is quite welcome to us. Here we shall quote two famous sayings.

As a working principle then we assume that the adoption of a specific verb : object order (i.e. VO or OV) in a language entails the modification of other syntactic characteristics, such as noun modifier order (Lehmann 1975: 156).

We can thus predict for any given language what its word order rules either are or will be, as soon as we know the order of object and verb in main clauses (Vennemann 1972/1974: 82).

We might think this assertion by Vennemann is superseded by his own later paper (1982). We are uncertain whether the maxim by Lehmann does not need reexamination and amendment. Such a prediction given in the quotations above would not be so easily admitted any more.

English may be said to have been a SVO language since the Middle English period (see Traugott (1972), Kohonen (1978), and Mitchell (1985)). As is evident in the third version of Vennemann's table, the English sentence construction is SV, that is, 'operator + operand' on one side, and VO, 'operand + operator' on the other. Thus English has both types AB and BA in one sentence. The preposition is an attribute of the BA language, and the structure of comparison also shows that English is a BA language. On the other hand, the order 'adjective + noun' is AB. This order may be said to be supported by another AB combination, that is, SV. This shows that in English the construction of SV (= AB) and that of VO (= BA) are struggling to exert influence over other linkings. We cannot but feel as if the order system of English were split and torn into two by the opposing powers. (7) We cannot say which order is the stronger.

Canale (1978) observes the disharmony between the verb phrase and the noun phrase concerning element order in English. He tries to offer various hypotheses to solve this problem, but only in vain. He cannot say anything more than Greenberg (1966), Lehmann (1972, 1974, 1975), and Vennemann (1972/1974). But now in Vennemann's third version of the table we have gained a knife to cut off the knot in question. At the same time the table (third version) will be the key to the question why the OV type still keeps productivity in noun compounding against the VO one.

As is well-known, Greenberg gives us two series of harmonious order of elements (Greenberg 1966: 100).

(a) Pr/ NG/VS/ VO/NA

(b) Po/ GN/SV/ OV/AN

In either series the indicated adposition (8) is harmonious with the following units in the same line. The former series indicates the typical order of the VSO languages (Type I languages), the latter series that of the SOV languages (Type III languages). The so-called Type II languages (SVO) cannot avoid being inconsistent in word order. We cannot formulate so harmonious a series for SVO languages as the former two. In English we have VO (= BA) on side, and SV (= AB) on the other. If we want to describe English word order with opposing trends in a similar way, it will be as follows.

(c) Pr/ VO : SV/ GN/ AN

The group of units before the colon and that after are disharmonious in order. It will be harder to make similar formulae for German and French. German has prepositions and postpositions. In French we meet with le petit garcon and les enfants terribles.

The compound carwash in the title of this paper has been said to display the construction of SOV order. As is said above, this order was dominant before the Old English period, and well preserved in subordinate clauses in Old English. In principal clauses the SVO order may be said to have been basic in Old English. Since the transitional period from Old English to Middle English, the SVO order has been predominant even in subordinate clauses. Concerning relative clauses and and-clauses, see Fujiwara (1986).

But this type of compounds has another important feature to be discussed, besides historical observations. Now we are going to examine the element order of deverbal compounds from the operator-operand point of view.

  (i) (Imper.)       Wash           a car.
                     V              O
                     operand        operator
                     B              A
 (ii) (Noun phrase)  car            washing
                     operator       operand
                     A              B
(iii) (Compound)     carwash(-ing)
                     operator       operand
                     A              B

The first thing to be noticed in these formulae is that the element wash is the operand, whether it is a verb or element of a phrase or compound. The element car is an operator whether it is the object of the verb or the first element of the compound or noun phrase. The phrase car-washing (noun or adjective) belongs to the AB pattern. In English the noun phrase is generally formed in the AB order, and noun compounding is in keeping with noun phrase construction. This order is supported by the SV construction, one of the two poles determining the word order system in a language of Type II. As is evident in the third version of Vennemann's table, the SV construction belongs to the AB pattern. It is the same with the construction of noun phrases, and so with that of noun compounds of the OV type.

Rent-a-car has not this support. If it is a verb phrase, rent a car, it realizes the usual BA order, but even when a noun phrase or compound. it has still a BA construction. The noun phrase is generally AB order even in Present-day English, unless a prepositional phrase or adjective clause is the element of A. The to-infinitive falls under the heading of the prepositional phrase. This gives the compound rent-a-car a little unconventional, and therefore even fresh feeling today. Now we can explain the following fact: against the group of window-cleaner and others cited in Lightfoot (1979: 160), pickpocket, breakfast, rent-a-car, etc. (VO type) are still in the minority of noun compounds.

As for the VS type of noun compounds, such as crybaby and jumpjet, it will not be so easy for the type to prevail. In the compounds cited above, cry and jump are operators, while in the sentences below the corresponding elements are operands.

The baby cries.

The jet jumps.

This duality, the present writer supposes, prevents these compounds from becoming prevalent.

Here I should like to express to Mrs. Yi, Hu Bok (Teikyo University) my deep sense of gratitude. Through her kind arrangement a copy of Vennemann (1982) was sent to me from Korea.

(1.) It is rather difficult to distinguish between compounds and phrases. Bauer (1983: 105) offers three requirements for an item to be regarded as a compound rather than as a syntactic phrase. They are (i) positional mobility, (ii) uninterruptability and (iii) internal stability. Carwash satisfies all these requirements. "Stress is not critical for compounds" (Bauer 1983: 104). But in [WND.sub.3] it has the most common stress pattern for compound nouns, if pronounced in isolation, that is, CARwash (if it is allowed to transcribe it after the method of Akmajian 1990).

(2.) According to OED, housekeep (vb.) was formed by back-formation from housekeeper or housekeeping. About this pattern of verb compounding (i.e. Noun + Verb) Bauer says, "the vast majority of this group arise from back-formation" (Bauer 1983: 208).

(3.) According to [COD.sub.9] an adder was formed from a naddre in ME by wrong division, that is, metanalysis. We may be permitted to apply the term to the process of dividing carwashing.

(4.) The term is used in Lehmann (1969, 1993). He sorts compounds into three classes, one of which is the synthetic compound. Akmajian uses the term in the same sense as Lehmann. In classification Bauer is different from those authors referred to above and a little more complex.

(5.) This paper was read by Vennemann at the Eleventh International Congress of Linguists, Bologna and Florence, in 1972 (Aug. 28-Sept. 2). He referred to this paper as Vennemann (1972) for a long time. But it is contained in the proceedings published in 1974. In this essay by the present writer it is indicated with Vennemann (1972/1974).

(6.) Lehmann (1986) criticizes Vennemann (1982) rather severely. But this paper by the present writer rests on Vennemann's assertion in his (1982) paper.

(7.) We might say as follows, too. Through word order change in the sentence, SOV altered to SVO. In other words, only the so-called VP underwent change, leaving the NP as it was, if we are allowed to use the terminology of the transformational grammar. The traditional element order of the NP is Operator + Operand. In carwash, wash is the head, that is, operand. In this way the noun phrase and noun compound keep the order of Noun + verbal (i.e. deverbal).

(8.) The term adposition is often used in typology. It covers both preposition and postposition. Cf. affix in morphology, which includes prefix, suffix and sometimes infix, too.

REFERENCES

Akmajian, Adrian et al.

1990 Linguistics -- An introduction to language and communication. Cambridge, Mass.: M.I.T. Press.

Anderson, John -- Charles Jones (eds.)

1974 Historical linguistics. Amsterdam: North Holland.

Bauer, Laurie

1983 English word-formation. Cambridge: CUP.

Canale, Michael

1976 "Implicational hierarchies of word order relationships", in: William J. Christie, Jr. (ed.), 39-69.

Christie, William J., Jr. (ed.)

1976 Current progress in historical linguistics. Amsterdam: North Holland.

van Coetsem, Frans -- Herbert L. Kufner (eds.)

1972 Toward a grammar of Proto-Germanic. Tubingen: Max Niemeyer.

Comrie, Bernard

1981 Language universals and linguistic typology. Oxford: Basil Blackwell.

Croft, William

1990 Typology and universals. Cambridge: CUP.

Fujiwara, Hiroshi

1986 "The relative clauses in Beowulf', in: Dieter Kastovsky -- Aleksander Szwedek (eds.), 311-316.

1995 "Landmarks of word order study (2)", Bulletin of the Department of English, Teikyo University 26: 183-221.

Givon, Talmy

1971 "Historical syntax and synchronic morphology", Publications of the Chicago Linguistic Society 7: 395-415.

Greenberg, Joseph H.

1966 "Some universals of grammar with particular reference to the order of meaningful elements", in: Joseph H. Greenberg (ed.), 73-113.

Greenberg, Joseph H. (ed.)

1966 Universals of language. Cambridge, Mass.: M.I.T. Press.

Heilmann, Luigi (ed.)

1974 Proceedings of the XIth International Congress of Linguistics. Bologna: Societa edifice il Mulino.

Kastovsky, Dieter -- Aleksander Szwedek (eds.)

1986 Linguistics across historical and geographical boundaries. Vol. 1. Berlin, New York: Mouton de Gruyter.

Kohonen, Vilja

1978 On the development of English word order in religious prose around 1000 and 1200 AD. Abe: Abe Akademi.

Lehmann, Winfred P.

1969 "Proto-Indo-European compounds in relation to other Proto-Indo-European syntactic patterns", Acta Linguistica Hafniensia 12: 1-20.

1972 "Proto-Germanic syntax", in: Frans van Coetsem - Herbert L. Kufner (eds.), 239-285.

1974 Proto-Indo-European syntax. Austin: University of Texas Press.

1975 "A discussion of compound and word order", in: Charles N. Li (ed.), 149-163.

1986 "Primes", in: Winfred P. Lehmann (ed.), 1-17.

1993 Theoretical bases of Indo-European linguistics. London, New York: Routledge.

Lehmann, Winfred P. (ed.)

1986 Language typology 1985: Papers from the Linguistic Typology Symposium, Moscow, 8-13 December, 1985. Amsterdam: Benjamins.

Li, Charles N. (ed.)

1975 Word order and word order change. Austin: University of Texas Press.

Lightfoot, David W.

1979 Principles of diachronic syntax. Cambridge: CUP.

Mitchell, Bruce

1985 Old English syntax. 2 vols. Oxford: Clarendon Press.

Traugott, Elizabeth C.

1972 A histoty of English syntax. New York: Holt, Rinehart and Winston.

Vennemann, Theo

1972/1974 "Analogy in generative grammar. The origin of word order", in: Luigi Heilmann (ed.), 79-83.

1974 "Topics subjects, and word order From SXV to SVX via TVX", in: John Anderson -- Charles Jones (eds.), 339-376.

1982 "What is a linguistic typology?", Studies in Linguistics 2: 1-42. (Taegu Linguistic Society, Korea.)

ABBREVIATIONS

[COD.sup.8] = The Concise Oxford Dictionary (8th edition.)

[COD.sup.9] = The Concise Oxford Dictionary (9th edition.)

RHD = The Random House Dictionary

WCD = Webster's New Collegiate Dictionary

[WID.sup.3] = Webster's Third New International Dictionary

[WND.sup.3] = Webster's New World Dictionary (3rd college edition.)

Po = Postposition

Pr = Preposition