Ditransitive alternation and theme passivization in Old English [1]

Tomohiro Yanagi
Chubu University

Abstract

1. Introduction

This paper is concerned with word order variation of Theme and Goal arguments in ditransitive passive constructions in Old English (OE). [2], [3] Here, Theme and Goal are two types of thematic roles often employed in the generative framework. The ‘Theme’ argument is roughly defined as an argument undergoing motion or a change of state; the ‘Goal’ argument refers to an argument receiving something in a situation. In OE ditransitive constructions, Theme and Goal arguments were marked with accusative and dative case, respectively; Goal arguments could either precede or follow Theme arguments without recourse to prepositions. [4] This is illustrated in (1).

The alternation in (1) contrasts with that observed in present-day English (PE). In PE, if a Goal argument follows a Theme argument, it must be accompanied with a preposition, as in (2).

In these active sentences, the grammatical functions of Theme and Goal arguments are direct and indirect objects, but when passivized, the objects adjacent to the main verb will be subjects as in (3) below. Thus, the terms Theme and Goal are employed instead of direct or indirect objects.

Furthermore, as indicated in (2) and (3), in PE, what is passivized is Goal in ditransitive constructions and it is Theme in to-dative constructions. In Standard English Theme cannot be passivized in ditransitive constructions, but it can in some British dialects, as in (4). This type of passive construction is called Theme passives. [5]

In OE, only Theme passives were possible, and such passives as in (3a) were not available because Goal arguments retained their dative case in passivization.

Let us now consider ditransitive alternation in OE. The pair of examples in (1) shows that two word orders, the ‘dative-accusative’ order and the ‘accusative-dative’ order, were observable in OE. Koopman (1990a), having looked at about one-fifth of the verbs listed in Visser (1963: §682), discussed these two word order patterns in main and subordinate clauses. [6] He then reached the conclusion that both orders are observed with the same frequency as summarized in Table 1. [7]

Given his conclusion, it would be expected that in passive constructions with ditransitive verbs, both the ‘dative-nominative’ and ‘nominative-dative’ orders are observable with close frequencies. But this expectation is not borne out, as will be shown in Section 2.

This paper is organized as follows: Section 2 classifies the word order patterns observed in ditransitive passive constructions of OE and provides their distributions and examples of each word order pattern. In Section 3 I will examine how ditransitive passive constructions are derived within the generative framework. The section further argues possible syntactic positions where the Theme and Goal argument occur in clause structure. Section 4 concludes the paper with a brief comment on the Latin influence on the word order patterns found in ditransitive passives.

2. (A)Symmetrical Word Order

This section provides examples of the two word order patterns in Theme passives. In order to collect relevant data, I used the York-Toronto-Helsinki Parsed Corpus of Old English Prose (Taylor et al. 2003; YCOE). [8] This is one of the descendants of the Helsinki Corpus and a ‘sister’ corpus of the Penn Corpora of Historical English. [9] The YCOE is syntactically annotated and those syntactic annotations as well as the POS tags were used for the retrieval of the data. For the retrieval CorpusSearch 2 was used as a searching program. [10]

The overall results are summarized in Table 2 below. This table shows the summary of the frequency of Theme passives classified according to the nominal/pronominal distinction of the arguments. Here, nominative noun and pronoun are Theme, and Dative noun and pronoun are Goal. The type of passive in which both arguments are pronouns is quite infrequent. The distribution is statistically significant at the level of significance (p < 0.01). [11]

In the following subsections we consider each of the two word order patterns: the ‘Theme-Goal’ order and the ‘Goal-Theme’ order.

2.1 The 'Theme-Goal' Order

The distribution of instances where Theme comes before Goal is summarized in Table 3. As we can see from Table 3, when the Goal argument is a noun, it follows the Theme noun or pronoun with little difference in frequencies; however, when the Goal is a pronoun, it tends to follow the Theme pronoun. The distribution shows no statistically significant difference.

Some of the actual examples taken from the YCOE are given in (5) through (12). Examples (5) and (6) are ones where Theme nouns precede Goal nouns. There are 74 instances of this combination found in the corpus.

Theme noun – Goal noun

In (5) se halga heap ‘the holy company’, which is a Theme, precedes þam hælende ‘the Saviour’, which is a Goal.

The number of examples where Theme nouns and Goal pronouns are involved is much smaller than the combination of Theme noun and Goal noun. I found only 17 instances. Here are two examples of this kind.

Theme noun – Goal pronoun

In example (7) and (8) the Theme nouns, þeos lar ‘this lore’ and ealle þing ‘all things’, are at the clause-initial position and the Goal pronoun, me ‘me’, occupies the position between the nominative subject and the finite auxiliary.

We go on to the sentences involving Theme pronouns. There are 75 occurrences of the Theme pronoun and Goal noun combination. It seems that whether the Theme argument is a noun or pronoun, it prefers to precede the Goal argument. Two examples are given in (9) and (10).

Theme pronoun – Goal noun

The last combination is the one in which both Theme and Goal are pronouns. Here are two examples out of the 30 instances.

Theme pronoun – Goal pronoun

In (11) the Goal pronoun him ‘him’ occurs between the two verbal elements, sceoldon ‘should’ and beon ‘be’. In (12) the pronouns, hi ‘they’ and me ‘me’, both occupy the pre-verbal position.

2.2 The 'Goal-Theme' Order

Let us next consider the ‘Goal-Theme’ order. The results obtained from the corpus research is summarized in Table 4. The total number is a little smaller than that of the ‘Theme-Goal’ order. It can be seen from Table 4 that the ‘Goal-Theme’ order is observed mainly when the Goal argument is a pronoun. However, the Goal pronoun is unlikely to precede the Theme pronoun.

Some examples of this order pattern are given in (13) through (19). The examples in (13) and (14) involve the Goal noun and Theme noun. This combination occurs 45 times in the corpus.

Goal noun - Theme noun

The Goal argument þisse cneorisse ‘this generation’ follows the auxiliary bið ‘be’ in (13), whereas in (14) the argument sumum men ‘some man’ precedes the auxiliary wæs ‘was’.

In contrast to the case of the Theme noun, the Theme pronoun can hardly follow the Goal noun. Only three instances were found. [12]

Goal noun – Theme pronoun

In this example the Goal noun eallum yflum ‘all evils’ occupies the clause-initial position and the Theme pronoun ic ‘I’ follows it and precedes the finite auxiliary.

It seems that the Goal pronoun can more easily precede the Theme noun than the Goal noun. The number of this type is more than twice as large as the one of the ‘noun-noun’ combination. Three examples are provided in (16)–(18).

Goal pronoun – Theme noun

In these examples, unlike most of the other examples, the Theme nouns seem to be at the clause-final position.

The last type, represented by only two examples, is the one in which both arguments are pronouns.

Goal pronoun – Theme pronoun

It is unlikely that the Theme pronoun follows the Goal noun or pronoun.

2.3 Summary

To sum up, I have qreorganized the tables presented in Sections 2.1 and 2.2 into Table 5 and 6.

In Table 5 the two word order patterns involving Theme nouns are classified according to the noun/pronoun distinction of the Goal argument. As we can see from this table, dative Goal nouns can either precede or follow nominative nouns though the ‘Theme-Goal’ order is more likely to occur (62.2% vs. 37.8%). On the other hand, dative pronouns almost always precede nominative nouns and they are less likely to follow nominative nouns (87.5% vs. 12.5%).

In Table 6, the calculation was made with the results involving Theme pronouns. Just as in those cases where dative pronouns occur before nominative nouns, as shown in Table 5, nominative Theme pronouns are much more likely to occur before dative nouns or pronouns. It thus can be said that the ‘nominative pronoun-dative’ order is the grammatical norm.

To summarize this section, we would have the following five points:

In Section 3 I will argue the derivation of each word order pattern within the generative framework.

3. Theoretical Discussion

In the previous section the relative word order of Theme and Goal arguments in ditransitive passive constructions was considered. This section discusses possible syntactic positions Theme or Goal arguments can occupy in clause structure, and suggests that both arguments can move to the specifier position of Tense Phrase (TP).

3.1 Derivation of Ditransitive Passivization

Before going to the main discussion, I will make a few assumptions relevant to the present analysis. First I follow Woolford (2006) in assuming tripartite Case distinction illustrated in (20). [14] Here, Nonstructural Case, i.e. Lexical and Inherent ones, are manifested in dative or genitive case in OE or rich inflectional languages such as Finnish. Finnish has much more case forms, 14 or 15 cases. These cases are closely related to semantic meanings. By contrast, nominative and accusative cases are structural ones. These cases are structurally licensed regardless of their semantic roles; the case forms are altered depending on their grammatical functions or structural positions.

(20)

Second, I assume the three-layered Verb Phrase structure in (21) for ditransitive verbs. Similar structures have been proposed in Ura (1996, 2000), Woolford (2006), and Yanagi (2011) as well.

(21)

In this structure va and vg assign the Agent and Goal role to the arguments in their specifier positions, respectively. The Theme role is assigned to the argument in the complement of VP.

A third assumption is related to case assignment. According to Woolford (2006), Nonstructural Cases are licensed as follows:

Adapting the assumptions in (22), I propose that the two objects of ditransitive verbs in OE are case-marked as in (23). [15] The Goal argument is case marked with dative inherently by vg, whereas the Theme argument is case marked with accusative structurally by va. It should be noted here that at the same time when the Goal argument is case marked by vg, it is also assigned the theta-role by the same head; on the other hand, the Theme argument is case marked by va, but it is assigned the theta-role by a different head V. [16]

(23)

Let us now consider ditransitive passivization. When a ditransitive verb is passivized, the functions of va will be inert. As a result, the Theme argument cannot be assigned accusative, and thus it is marked as nominative by T. The functions of vg are still active after passivization, by contrast, the case assigned to the Goal argument is retained as dative. This is schematically illustrated in (24).

(24)

3.2 Syntactic Positions of Theme and Goal

So now we are in the position to discuss syntactic positions of Theme and Goal in clause structure. After the derivation reached (24), the structure merges with a functional category, C(omplementizer), which projects CP. This yields (25). [17]

(25)

Given that OE is a Verb-Second language, finite verbs move up to C through T in main clauses. In passive sentences the auxiliary BE (beon, wesan, ‘be’ or weorþan ‘become’) moves up to C. In addition, the specifier position of CP must be filled with an element, such as an adverb and a nominal. Let us take example (5), repeated here as (26), for illustration purpose.

As mentioned above, the auxiliary wearð ‘was’ occupies the head of the CP after moving through T. The adverb þa ‘then’ occupies the specifier position of the CP. The initial part of the structure would be like (27). [18]

(27)

The general assumption is that T has the specifier position to be filled with a nominal element. This is called the EPP (Extended Projection Principle) requirement (cf. Chomsky 1981, 1995, 2001). The element filling the position is usually a noun phrase marked with nominative. In the case of (26) se halga heap ‘the holy company’ is located at this position to fulfill the EPP requirement. The nominative noun phrase se halga heap ‘the holy company’ moved up there from the complement of VP, i.e. the position of DPtheme in (25). Then we have the following structure:

(28)

The remaining two constituents, þam hælende ‘the Saviour’ and geoffrod ‘offered’, get stuck at their base-generated positions unless they are required to move for syntactic or semantic reasons. Therefore, the whole sentence would have the structure in (29). This is a typical structure of the ‘Theme-Goal’ order.

(29)

In a V2 sentence, the initial element is not limited to an adverb, as in (26). Instead, a noun phrase marked with dative can occupy the specifier position of the CP. Now we consider this case, taking (14), repeated here as (30), as an example.

Provided that the specifier of the TP is occupied by a nominative nominal element, as in the case of (26), the structure of (30) would be like (31), which is a typical structure of the ‘Goal-Theme’ order in passive sentences.

(31)

In this structure the Theme argument unlybba ‘poison’ is in the specifier position of the TP to fulfill the EPP requirement; the Goal argument sumum men ‘some men’ is topicalized into the clause-initial position.

So far we have proposed two different structures for the two word order patterns: (29) for the ‘Theme-Goal’ order and (31) for the ‘Goal-Theme’ order. A language blueprint, however, is not so simple. When one of the two arguments in a passive sentence is a pronoun, it possibly occupies a different position from the one where a noun phrase does. The two relevant examples are repeated here as (32) and (33).

In (32) and (33) the Goal and the Theme arguments are pronouns, respectively. It should be noted here that the pronouns in both examples occur between the initial element and the finite auxiliary. These cases can be treated as cliticization. [19] The pronouns are adjoined to the head of CP, as schematically illustrated in (34) and (35).

(34)

(35)

With respect to the syntactic position, these pronouns are located at the same position, irrespective to their case.

Furthermore, there seems to be a general tendency in the order of pronouns: a subject pronoun tends to precede an object one. [20] This can explain why the frequency of the ‘Goal-Theme’ order is quite low when the two arguments are pronouns (cf. Table 4).

Finally, let us consider example (36), repeated from (13)

This, like (30), is an instance of the ‘Goal-Theme’ order, but the Goal argument is not located at the clause-initial position. The initial element is the negative particle ne and the auxiliary occupies the head of CP. [21]

Given the structure in (24), it can be said that in (36) the noun tacen ‘token’ is assigned nominative case by T at the base-generated position and gets stuck there. Here a question arises as to the position of the Goal argument þisse cneorisse ‘this generation’. A possible answer is that the argument stays at the base-generated position. However, this might cause a violation of the EPP requirement: this requirement cannot be fulfilled if no element moves into the specifier position of the TP. Therefore, in order to avoid this problem, I propose that the Goal argument in (36) occupies the specifier position of the TP. There needs to be close scrutiny to this proposal, but this is left open for future theoretical research.

4. Concluding Remarks

Before concluding this paper, I will make a brief comment on Latin influence on OE ditransitive passive constructions. For comparison, I picked up the subcorpora in which both word order patterns are observed and the number of each pattern is 5 or more. Such subcorpora are provided in Table 7.

As an overall tendency, in translations from Latin the ‘dative-nominative’ or the ‘Goal-Theme’ order is preferred whereas in OE originals the ‘nominative-dative’ or the ‘Theme-Goal’ order is preferred. [22] This might be due to the high frequency of use of dative pronouns. If the data is restricted to nouns, we would have the results in Table 8. Table 8 consists of the same subcorpora in Table 7, irrespective of the number of the tokens, for ease of comparison.

From this table we can see that the ‘Theme-Goal’ order is more frequent than the ‘Goal-Theme’ order only in Ælfric’s three works: Catholic Homilies, First and Second series, and Lives of Saints. Since the total number of the occurrences is small, it might be difficult to make a definitive conclusion, but it could be said that the ‘Theme-Goal’ order is preferred in OE unless pronouns are used. [23]

In this paper it was shown that when the Theme and Goal arguments are both nominal in passive sentences of OE, the ‘Theme-Goal’ order is more frequent than the ‘Goal-Theme’ order, and that when they are both pronominal, the ‘Goal-Theme’ order is quite rare. The former fact seems not to be compatible with the observation of Koopman (1990a), who suggests that both word order patterns are almost the same in frequency. As for the latter fact, it was proposed that this rarity is due to the general tendency of word order of clitic pronouns. I further argued the syntactic positions of the Theme and Goal arguments. In normal main clauses, Theme and Goal occupy the specifier position of TP and the base-generated position, respectively, yielding the ‘Theme-Goal’ order. When a dative nominal is topicalized, however, it occupies the specifier position of CP and the word order is altered to the ‘Goal-Theme’ order. Furthermore, if one of the two arguments is a pronoun, it is cliticized to the head of CP.

Notes

[1] An earlier version of this paper was presented at the Helsinki Corpus Festival: The Past, Present, and Future of English Historical Corpora, held at Tieteiden talo, Helsinki, Finland, on 28 September – 2 October 2011. I am very grateful to the audience and anonymous reviewers for their invaluable comments and suggestions. All remaining errors and inadequacies are my own. This research is supported in part by JSPS Grant-in-Aid Scientific Research (C), No. 23520598.

[2] As pointed out by an anonymous reviewer, Theme and Goal might convey a different meaning depending on a linguistic theory. For example, ‘theme’ as well as ‘rheme’ are terms in the Prague School, and taken over in the pragmatic sentence analysis. The ‘theme’ expresses old, familiar information, while the ‘theme’ expresses new information.

[3] In some cases Benefactive may be more suitable than Goal to refer to indirect objects. However, I use Goal in a broader sense to cover both Goal and Benefactive arguments for the reason of simplicity; this semantic distinction does not affect the present analysis.

[4] There were a few other combinations of cases observed: accusative and accusative; accusative and genitive. Here are two examples:

In (i) hiene ‘him’ is accusative and fultumes ‘help’ is genitive; in (ii) him ‘him’ is dative and þæs rices ‘the kingdom’ is genitive. Examples of this kind are not dealt with in this paper.

[5] See Haddican (2010) for detailed discussion of Theme Passives in British English dialects.

[6] Koopman does not mention the verbs consulted in his study, but Visser’s list contains bringan ‘bring’, geoffrian ‘offer’, and sillan ‘give’. These ditransitive verbs are also found in the passive constructions under consideration. Koopman took the examples from A Microfiche Concordance to Old English (Healey and Venezky 1980).

[7] For relevant discussion of Middle English counterparts and diachronic perspectives see Allen (1995), Denison (1993), McFadden (2002) and Polo (2002).

[8] The manual of the YCOE is available at <http://www-users.york.ac.uk/~lang22/YcoeHome1.htm>.

[9] For more information of the Penn Corpora of Historical English, visit <http://www.ling.upenn.edu/hist-corpora/>.

[10] This program can be downloaded at <http://corpussearch.sourceforge.net/>.

[11] For the present analyses JavaScript-STAR version 5.5.7j (2010.12.06) was used to calculate the χ² values. This Java-based calculator was created by Hiroyuki Nakano, and is workable at <http://www.kisnet.or.jp/nappa/software/star/>.

[12] This ‘deviant’ word order may be intriguing. Its rarity might be due to pronominalization of the Theme argument, but close scrutiny will be needed to determine how this word order is derived.

[13] This may naturally follow from the general observation that given information tends to precede new information. Since pronouns usually represent given information, they are likely to precede nominal elements. From a generative point of view, the given information may be treated as a driving force of movement.

[14] It has been proposed in the generative literature that argument noun phrases must be assigned case in any form. Case assignment procedures have been ‘refined’ in the history of the generative grammar (cf. Chomsky 1981, 1995, 2001).

[15] In (23) the detailed mechanism of case marking is put aside.

[16] The Agent argument is assigned the theta-role by va and it is case marked with nominative structurally by T. For this analysis see Yanagi (2011). See also Hiraiwa (2001) for a different approach.

[17] For the clause structure of OE see also Kemenade (1987), Pintzuk (1999) and Roberts (1993) among others.

[18] An actual derivation is not counter-cyclic, but it proceeds in a bottom-up fashion. Here, only for expository purposes, the derivation process is described acyclically.

[19] For cliticization in OE see Kemenade (1987), Pintzuk (1996, 1999) and Koopman (1990b), among others. See also Yanagi (2000) for a theoretical approach to cliticization.

[20] In general, the word order of clitic pronouns seems to be fixed in a given language. For French, a clitic language, see Jones (1996).

[21] The clause in question is embedded in the complement of secgan ‘say’. Normally embedded clauses do not exhibit Verb-Second phenomena, but they do sometimes when they are complements of verbs of saying, as in (36).

[22] This does not necessarily mean that the ‘Goal-Theme’ order is preferred in Latin, but this point is left open.

[23] Ælfric's compilations often had direct sources in Latin homilies, sermons and exegesis, which may have affected his OE usage. I am grateful to Olga Timofeeva for this point.

References

Allen, Cynthia. 1995. Case Marking and Reanalysis. Oxford: Oxford University Press.

Chomsky, Noam. 1981. Lectures on Government and Binding. Dordrecht: Foris.

Chomsky, Noam. 1995. The Minimalist Program. Cambridge, Miss.: MIT Press.

Chomsky, Noam. 2001. “Derivation by Phase”. Ken Hale: A Life in Language, ed. by Michael Kenstowicz, 1–52. Cambridge, Miss.: MIT Press.

Denison, David. 1993. English Historical Syntax. London: Longman.

Haddican, William. 2010. “Theme-Goal Ditransitives and Theme Passivisation in British English Dialects”. Lingua 120: 2424–2443.

Healey, Antonette DiPaolo & Richard L. Venezky. 1980. A Microfiche Concordance to Old English. Toronto: Pontifical Institute of Mediaeval Studies.

Hiraiwa, Ken. 2001. “Multiple Agree and the Defective Intervention Constraint in Japanese”. Proceedings of the HUMIT 2000 (MIT Working Papers in Linguistics 40): 67–80.

Jones, Michael Allan. 1996. Foundations of French Syntax. Cambridge: Cambridge University Press.

Kemenade, Ans van. 1987. Syntactic Case and Morphological Case in the History of English. Dordrecht: Foris.

Koopman, Willem. 1990a. “The Double Object Construction in Old English”. Papers from the 5th International Conference on English Historical Linguistics, ed. by Sylvia Adamson et al., 225–243. Amsterdam: John Benjamins.

Koopman, Willem. 1990b. Word Order in Old English: With Special Reference to the Verb Raising. Ph.D. dissertation, University of Amsterdam.

McFadden, Thomas. 2002. “The Rise of the To-Dative in Middle English”. Syntactic Effects of Morphological Change, ed. by David W. Lightfoot, 107–123. Oxford: Oxford University Press.

Ono, Shigeru and Tosho Nakao. 1980. Eigoshi I [History of English I]. Tokyo: Taishukan.

Pintzuk, Susan. 1996. “Cliticization in Old English”. Approaching Second: Second Position Clitics and Related Phenomena, ed. by Aaron L. Halpern & Arnold M. Zwicky, 385–49, Stanford: CSLI.

Pintzuk, Susan. 1999. Phrase Structures in Competition. New York: Garland.

Polo, Chiara. 2002. “Double Objects and Morphological Triggers for Syntactic Case”. Syntactic Effects of Morphological Change, ed. by David W. Lightfoot, 124–142. Oxford: Oxford University Press.

Roberts, Ian. 1993. Verbs and Diachronic Syntax: A Comparative History of English and French. Dordrecht: Kluwer.

Taylor, Ann, Anthony Warner, Susan Pintzuk, & Frank Beths. 2003. The York-Toronto-Helsinki Parsed Corpus of Old English Prose. York: University of York.

Ura, Hiroyuki. 1996. Multiple Feature-Checking: A Theory of Grammatical Function Splitting. Ph.D. dissertation, MIT.

Ura, Hiroyuki. 2000. Checking Theory and Grammatical Functions in Universal Grammar. Oxford: Oxford University Press.

Visser, Frederik Theodoor. 1963. An Historical Syntax of the English Language. Part I. Leiden: Brill.

Woolford, Ellen. 2006. “Lexical Case, Inherent Case, and Argument Structure”. Linguistic Inquiry 37: 111–130.

Yanagi, Tomohiro. 2000. “Some Notes on Cliticization: Attract or Greed”. IVY 33: 31–59.

Yanagi, Tomohiro. 2011. “Case Licensing of Noun Phrases in the History of English”. JELS 28: 176–182.