On the development of a prepositional object construction with give verbs, motion verbs and Latinate verbs in English

Katarzyna Sówka-Pietraszewska
University of Wrocław

Abstract

1. Double Object Constructions in Old English and early Middle English

Introduction

It is a very well-known fact that dative alternation is not a creation of the OE era. The historical approaches to double object construction emphasize that it was not until the late OE and early ME period that verbs with an inherent meaning of possession change started to show two realizations of their arguments (McFadden 2002, Mitchell 1985). [1] In OE, a double object construction (DOC) with two nominal objects [V NP1 NP2] significantly prevailed in number over the structure with an oblique object [V NP2 to-NP1] (henceforth called a prepositional object construction (POC)); for DOC see (1a), for POC (1b).

POC was restricted to the verbs of speaking (e.g., cweþan, sprecan) and was selected by motion verbs (e.g., bringan, lætan, niman, sendan), however, far less frequently than DOC.

The number of POCs started togrow in the early ME period. This change involved not only motion verbs but, following McFadden (2002), also verbs with an inherent meaning of possession change with no implied movement, which started to select POC. [2] [3] The example in (2) shows one of the first occurrences of give in POC with a recipient, found in HC, dated from the early ME period.

In this paper, I adopt McFadden’s (2002) claim that the construction [V NP2 to-NP1] with give in (2) was a recently emerged construction. In light of this statement, I assume that the similarity of (1b) and (2) is only a surface similarity. Since the verbs send and give lexicalize contrasting inherent meanings (caused motion in (1b) and caused possession in (2)) each variant should be associated with the respective meaning. This statement is supported during the further analysis.

As the semantic encoding of constructions in (1a) and (2) has not been subject to any formal study yet, the main concern of this paper is first to examine the meaning associated with each variant and second, using the data from this analysis, to motivate the puzzling structure selection of Latinate verbs in early ModE and in present day English. The structure selection of Latinate verbs has been the subject of heated debate (Grimshaw 2005, Grimshaw and Prince 1986, Krifka 2004, Pinker 1989) because these verbs realize their arguments only in POC, regardless of the fact that some of them do not involve caused motion. As a result, such Latinate verbs as donate and contribute occur in the structure with, puzzlingly, a direction-pointing preposition. However, historical data show that their structure selection is not without support.

This paper (inspired by Rappaport Hovav and Levin’s (2008) analysis for ModE) examines the variants of POC meaning in contexts with ‘give’ verbs and ‘motion’ verbs in the history of English. The main claim in this paper is that the analysis of POC meaning was conditional on the meaning lexicalized by verbs: (i) POC with ‘give’ verbs, as in (2), realized a caused possession meaning, and (ii) POC with ‘motion’ verbs, as in (1b), realized a caused motion meaning that resulted from the intention of possession change. Consequently, the occurrence of give verbs in POC in ModE should not result in them being analyzed as involving caused motion. This approach will motivate this claim historically.

This paper is organised as follows: Section 2 shows the development of POC in OE and ME and proposes a hypothesis assuming the coexistence of two variants of POC in the ME period. The following section discusses the distinct roles of the preposition to in both of these constructions. The event schemas for these constructions are presented in Section 3. The last analytical section contains a historical analysis of Latinate verbs.

The database for this analysis includes the verbs that belong to the following groups: (i) possession change verbs with no implied movement, which include the predecessors of the most characteristic successful possession change verbs: give and grant, henceforth called ‘give’ verbs, (ii) possession change verbs involving movement: predecessors of bring and send in respective periods, henceforth called ‘motion’ verbs, and (iii) Latinate verbs: donate, report, explain, distribute, illustrate, recite, transport, and contribute, from which donate, and contribute were selected into the group of Latinate ‘give’ verbs. The data for this analysis comes from Helsinki Corpus TEI XML Edition, first edition (2011) (HC). A few examples come from the sources outside HC but they are annotated respectively.

2. Structure selection of give verbs and motion verbs in OE and early ME

It appears to be the case that in OE the set of ‘give’ verbs was not very abundant and they were almost all of purely Germanic origin. The most representative examples of ‘give’ verbs were: giefan, agiefan, sellan, tiþian, unnan, and gebōcian. ‘Give’ verbs selected DOC in which the indirect object (IO) was realized as the recipient. The IO either immediately followed the verb [V IO DO] or occurred following the direct object [V DO IO]. Both argument orders were equally popular: 54% [V IO DO] and 46% [V DO IO] (Koopman 1993: 112). [4]

Interestingly, the data from the corpus contains the following example of give in POC , see the example in (3) below. [5]

In the OE period POC with ‘give’ verbs was scarce, and was outnumbered by DOC. Only one example of POC with give was found in HC. The low frequency of give verbs in POC in OE is further supported in Mitchell (1985, p.512, §1210), who admits to having found several such constructions, for example with agiefan followed by prepositional monastery IO, and place IO or (ge)sellan followed by prepositional IO a church. At first, the semantic role of these IOs seems to be an inanimate goal; however, on second thoughts they turn out to be animate because they represent organizations of people. These people come into possession of the given goods. Consequently, they should be assigned the semantic role of recipients. [6]

The example of give in a construction which involved a preposition with the main sense of ‘to’ or ‘alongside’ is rather intriguing. However, the low number of such constructions and the restricted type of recipients (no datives of person) suggest that they might have been created on analogy with constructions involving ‘motion’ verbs, for example send. In OE and early ME send more often occurred in POC than in DOC when it was followed by a goal, see (4).

The structure like in (4) above will be further called a directional POC. It is assumed that since it involved an allative preposition it was associated with caused motion and emphasized an intended possession transfer, rather than a successful possession change.

The popularity of POC with personal recipient started to grow in the early ME. According to McFadden (2002: 111), POC with a give verb was a construction that arose from DOC [V DO IO], i.e., DOC [V DO IO] > POC [V DO to-IO], to remedy for misinterpretations following the loss of the morphological case. Given that POC with ‘give’ verbs followed by person recipient was a new construction and that it developed from DOC, it follows that it accompanied the directional POC, which had already been present in the OE period and was discussed above.

The popularity of the new POC grew slowly over the course of the ME period. McFadden shows that at the beginning of the ME period the popularity of this construction was rather low. Only 7.9% of all double object constructions with ‘give’ verbs found in texts from the period 1150–1250 are POCs. This growth accelerates in the period between 1250–1350, when their number rises rapidly. In Ayenbite of Inwyt almost 35% of the constructions with ‘give’ verbs are POCs (McFadden 2002: 111).

As a result of the growing popularity of POCs with ‘give’ verbs, DOCs with the argument order [V DO IO] disappeared entirely. It seems that this growth in the popularity of POC with give verbs was conditional on the pace of withdrawal of the morphological case and consequently of the construction [V DO IO] (see McFadden 2002 for detailed discussion). On the other hand, the directional POC was also used more frequently. For example in the O3 period (950–1050), only one of three double object constructions with bringan was POC, but in the M2 period (1250–1350) the number of POCs equaled the number of DOCs. In contrast, sendan, which in the O4 period (1050–1150) occurred mainly in POCs, with only one exception for a DOC, in M1 (1150–1250) and M2 (1250–1350) started to be accompanied by DOC, i.e., the M1 period: 9 DOCs and 11 POCs; and in the M2 period: 9 DOCs and 9 POCs.

Summarizing this analysis, let us propose a claim, which will be tested in the next sections of this paper.

Claim: In the early ME period, POC realized two types of the inherent meaning of double object verbs. The variants were: the caused possession variant for ‘give’ verbs and the caused motion variant for ‘motion’ verbs. Different underlying structures of POC derived from different meanings lexicalized by verbs and were additionally supported by disparate roles of the preposition to in the variants.

The assumption concerning distinct roles of the preposition to postulates that in the directional POC the preposition to was a direction pointer, whereas in POC, which was a converted DOC, the preposition was a grammaticalized dative case marker (this construction will be henceforth called a dative POC).

Let us analyze the role of the preposition in the dative POC and provide a more formal discussion of the semantic role of the preposition to in ME.

3. The role of the preposition to in the dative POC and the directional POC

It has been observed that the use of the same preposition in POCs made the structures look overtly identical. In order to show the semantic difference, let us outline the structural case marking and the semantic encoding in IO, first in OE DOC, next in the dative POC, and finally in the directional POC. I will sketch IO using the Case Phrase (KP) approach from Schoorlemmer (1994). [7]

In OE DOC [V DO IO], the IO involved a head K, which selected a complement NP with a theta-role of recipient. This theta-role was compatible with the dative’s directional semantics.

(5)	IO in DOC

During the process of the overt case marking withdrawal, KP could not be projected as the dative case was lost. In order to avoid misinterpretation, the dative case-marking preposition to was introduced to the head position. To selected a complement NP with a theta-role of a recipient.

(6)	IO in the dative POC (early ME)

The previous section claimed that at the beginning of the ME period, the dative POC was not the only type of POC, but it coexisted with the directional POC. In the directional POC, the preposition to had only a role of a direction pointer, see (7).

(7)	IO in the directional POC (OE and early ME)

Having completed the presentation of the different functions of the preposition in POCs, let us now analyze the semantic event structures of POCs.

4. Meanings lexicalized in the directional POC and the dative POC

As was shown above, the directional POC was already present in OE and was selected mainly by ‘motion’ verbs. In Section 2 it was moreover assumed that the use of the allative preposition implied that the focus of this construction was an intended possession change or transfer but a rather unsuccessful one. The preposition to was assumed to have been the marker for the path of transfer towards the goal or the recipient. Given this observation, I assume that the events realized in the directional POC can be schematically summarized as in (8) below.

By contrast, the dative POC realized meanings of lexically different verbs. Recall that the dative POC was a converted OE DOC. In Section 2 it was assumed that OE DOC was selected by verbs lexicalizing possession change, with no implication on movement. By analogy, the semantic events realized in the dative POC must be the same as those in OE DOC. A dative marking preposition to strengthened the encoding of successful possession change.

In summary, during the course of ME the alternating verbs realized their arguments in the following structures: (i) give verbs in DOC and the dative POC, and (ii) motion verbs in DOC and the directional POC.

The second aim of this paper was to analyse the puzzling structure selection of Latinate verbs. These verbs, which were adopted into English from Latin mainly via French, disallow DOC in ModE. Intriguingly, some of them, for example donate or contribute, are synonymous to give; thus one could expect them to occur in DOC. This observation raises the following questions: (i) in which structure or structures did Latinate verbs realize their inherent meaning shortly after being introduced in English, (ii) assuming that they occurred in POC, in which of the two available types of POC did they occur, and (iii) why do give verbs from this group reject DOC? Let us try to provide the answers to these questions in the next section. I will search for the answer by analyzing the Latinate verb data of the late ME and early ModE.

5. Latinate double object verbs in the ME and ModE periods

As stated above, the main puzzle about Latinate double object verbs is their structure selection, which at times seems not to be in complete harmony with the inherent meaning of these verbs. This lack of semantic meaning agreement is particularly striking with Latinate ‘give’ verbs, such as donate or contribute, which occur in POC, but do not take DOC. However, the data from ME show that this is not a hard-and-fast rule. Latinate verbs did not always reject DOC. On the contrary, during the period between the late 14th and 17th centuries, some of the recently introduced Latinate verbs, e.g., deliver, declare, concede, and contribute occurred in DOC, see (10) below.

The examples above show that in late ME and early ModE Latinate verbs could select DOC, although POC was also fully productive in this period. It is important to point out that DOC did not occur regularly. Data from HC show that over the period E1-E3 (1500–1710), report, distribute, recite, transport and contribute, all verbs found in ditransitive use, occurred exclusively in POC. [9] It seems that the selection DOC was rather spontaneous.

Visser (1963: 624, §687) additionally observes that over the course of the 14th and 15th centuries, the overall number of structures with to-IO increased significantly. The construction with to-IO was favoured even in contexts which in ModE would be considered ungrammatical, see (11) below.

The chaotic use of to-IO became restricted after 1500 (Visser 1963: 624, §687), but Latinate verbs started to show more regular structure selection later. The data from Visser (in (10)) show the latest example of Latinate verbs occurring in DOC in the 17th century, hence in the early ModE period. What was the reason for an eventually unified structure selection?

The observation that Latinate verbs, shortly after being introduced into English, realized their arguments in DOC and POC, but later started to avoid DOC, raises some questions concerning the inherent semantic encoding of these verbs. Should it be assumed that the inherent semantic encoding of Latinate ‘give’ verbs has been modified over the years? Did the structure selection of ‘give’ verbs, e.g., contribute (10d), became restricted to POC in ModE because their inherent meaning of caused possession has changed into the caused motion meaning? This claim does not sound plausible; therefore I need to search for factors outside the scope of semantics.

So far, few suggestions have been made towards accounting for the exclusive selection of POCs by Latinate verbs. One such suggestion is the ‘semantic complexity’ hypothesis by Pinker (1989: 123), which assumes that Latinate verbs are more complex semantically than their Anglo-Saxon synonyms, e.g., donate as opposed to give. However, it remains unclear why the semantic complexity of a verb should disqualify DOC. Krifka (2004: 5) examines the syntactic influence hypothesis, which assumes that the structure selection of borrowed verbs could have been influenced by the source language. French does not have a construction with two bare NP objects [V NP NP], but only POC [V NP à-NP]. However, Krifka concludes that the origin of the verb should not influence its structure selection in the target language. Grimshaw (2005) argues for a phonological approach to Latinate verbs, in which the metrical criterion plays a crucial role in structure selection. Accordingly, she divides verbs into two groups: (i) verbs with one metrical foot (monosyllabic verbs, verbs with an initial stress, and verbs with second-syllable stress if the first syllable is a schwa); and (ii) verbs with two metrical feet. The verbs from the first group alternate, in contrast to the verbs from the second group, which are non-alternating verbs. Interestingly, all Latinate verbs involve more than one metrical foot, for example donate, report and explain contain two syllables, and the second syllable bears primary stress; this means that they have two metrical feet. As a result, their structure selection is restricted only to POC. On the other hand, offer and promise, which are Latin borrowings, have initial stress and belong to the group of verbs whose prosodic weight is no more than one foot. This feature makes them participate in dative alternation.

Hence, it seems convincing that it was the phonological constraint, which limited the choice of the syntactic structures for Latinate verbs. However, the phonological approach does not contradict the findings of this analysis. The previous sections have assumed the coexistence of two semantically distinct types of POC for the realization of the arguments of double object verbs. I conclude that Latinate ‘give’ verbs cannot realize their arguments in DOC due to the phonological constraint discussed above, but they may lexicalize the caused possession meaning in the dative POC. In contrast, Latinate ‘motion’ verbs lexicalize their caused motion meaning in the directional POC, see the schemes in (12) below.

This evidence from the Latinate verbs additionally supports the hypothesis assuming two underlying variants of POC meaning.

6. Conclusion

The main aims of this paper were, first, to examine the variants of POC meaning in OE and ME and second to motivate the puzzling structure selection of Latinate verbs in late ME, early ModE and present-day English.

The historical semantic analysis of the constructions with ‘give’ verbs in Section 2 showed that they licensed two ways of expressing their arguments. In the OE period they occurred in DOC with two nominal objects arranged in a relatively free order. In this time span DOC significantly prevailed over POC. However, in ME ‘give’ verbs occurred in DOC [V IO DO] along with POC [V DO to-IO]. Following McFadden (2002) it was assumed that this POC arose in early ME from DOC [V DO IO] to prevent the ambiguities after the loss of the morphological case. It was moreover assumed that, after being introduced, this so called dative POC (with a dative case marking preposition) occurred alongside with the directional POC, which was already present in OE with ‘motion’ verbs. This observation led to the main hypothesis of this paper, which claimed that in the ME period POC realized different types of verbs meaning, as a result, two main types of POC were distinguished: (i) the directional POC which was selected by ‘motion’ verbs and (ii) dative POC selected by ‘give’ verbs. Section 3 characterized the functions of the preposition to in the variants. Respectively, in the directional POC the function of the preposition was a direction pointer. In dative POC, on the other hand, to was a grammaticalized preposition, whose role was restricted to dative case marking. In Section 4, the discussion gave rise to the following observation concerning the verbs meaning that were be realized in POC: (i) the directional POC: realized the meaning of caused motion resulting from the intention of possession change and (ii) the dative POC: realized the meaning of caused possession.

This observation of two variants of verbs meaning realized by POC, support Rappaport Hovav and Levin’s (2008) ‘verb-sensitive’ approach to ModE alternating verbs. In this approach, ‘give’ verbs have only a caused possession analysis in POC, while ‘motion’ verbs convey a caused motion or a caused possession reading. In light of this approach, the following ModE sentences: John gave Mary flowers. and John gave flowers to Mary. have only one and the same semantic analysis of caused possession. Interestingly, historical data has shown that POC selected by ‘give’ verbs in the early ME period was a construction with the same semantic encoding as in OE DOC. A conclusion reached from this observation states that although ‘give’ verbs select the same construction as ‘motion’ verbs (for example send or throw) this fact does not classify them as ‘motion’ verbs, because it is the dative variant that realizes the inherent meaning of caused possession lexicalized by ‘give’ verbs. In this paper, this fact is claimed to be historically conditioned.

The second part of the analysis, in section 5, tackled the problematic structure selection of Latinate verbs, but in particular of Latinate ‘give’ verbs. Interestingly, the late ME and early ModE data showed that Latinate verbs sometimes realized their arguments in DOC. The event structure of DOC was said to have had the most compatible semantic encoding with Latinate ‘give’ verbs. However, it was shown that in course of ModE the metrical complexity (Grimshaw 2005) of these verbs overshadowed their semantic properties and eventually led to the unanimous selection of POC. This change, however, did not influence the inherent meaning lexicalized by Latinate ‘give’ verbs. Following the results of this analysis, it was assumed that since POC replaced DOC in all contexts, Latinate ‘give’ verbs realize the caused possession meaning in the dative POC. On the other hand, Latinate ‘motion’ verbs select the directional POC.

In summary, this approach shows that ‘give’ verbs in POC should not be analyzed as involving caused motion in their inherent meaning.

Notes

[1] See, among others, McLaughlin (1983), Koopman (1990), (1993), (1998), Allen (1995), Weerman (1997), Kroch and Taylor (2000), Polo (2002), Denison (2004) for an overview and general discussion of the double object constructions in OE.

[2] The database for the analysis in McFadden (2002) includes the verbs give, grant, show, teach and tell, which he calls true dative verbs.

[3] Possession change verbs with no implied movement include verbs with the inherent meaning of successful possession change. In this paper, whenever possession change verbs with no implied movement are mentioned, the term refers to ‘give’ verbs, i.e., give and grant.

[4] See McFadden (2002) for a detailed discussion on surface word orders in ME double object constructions.

[5] I would like to thank Anastasia Eseleva for pointing out the example with give in POC to me, and for the discussion.

[6] This observation was inspired by a discussion with Matti Kilpiö.

[7] KP (Case Phrase) was introduced by Schoorlemmer (1994), who assumed that the dative case consists of functional projection KP. In KP, the head carries the functional value “DAT”, whose specifier carries the semantics common to all ‘regular’ datives, which he calls directional. K selects a complement NP with a theta-role (RECIPIENT, GOAL or EXPERIENCER) that is compatible with the dative’s directional semantics. It can also take a complement NP without a theta-role, which results in a dative adjunct (Schoorlemmer 1994: 136).

[8] The semantic schemes of POC in (8) and (9) are modified neo-Davidsonian representations from Krifka (1999).

[9] Example (10d) was not found in HC and is not included in the analysis.

Corpora

Helsinki Corpus TEI XML Edition. (2011). First edition. Designed by Alpo Honkapohja, Samuli Kaislaniemi, Henri Kauhanen, Matti Kilpiö, Ville Marttila, Terttu Nevalainen, Arja Nurmi, Matti Rissanen and Jukka Tyrkkö. Implemented by Henri Kauhanen and Ville Marttila. Based on The Helsinki Corpus of English Texts (1991). Helsinki: The Research Unit for Variation, Contacts and Change in English (VARIENG), University of Helsinki.

Taylor, A., A. Warner, S. Pintzuk, and F. Beths (2003). The York-Toronto-Helsinki Parsed Corpus of Old English Prose. Helsington, York: University of York.

The British National Corpus, version 3 (BNC XML Edition). (2007). Distributed by Oxford University Computing Services on behalf of the BNC Consortium. URL: http://www.natcorp.ox.ac.uk/

References

Allen, C. L. 1995. Case Marking and Reanalysis: Grammatical Relations from Old to Early Modern English. Oxford: Oxford University Press.

Denison, D. 2004. English Historical Syntax: Verbal Constructions. Longman Linguistics Library. London: Longman.

Grimshaw, J. 2005. “Datives, Feet and Lexicons”. Words and Structure, CSLI Lecture Notes, no. 151, 107–128. University of Chicago Press.

Grimshaw, J. & A. Prince. 1986. A Prosodic Account of the to-Dative Alternation. Unpublished manuscript, Brandeis University.

Green, G. 1974. Semantics and Syntactic Regularity. Bloomington, Ind.: Indiana University Press.

Hale, K. & S. J. Keyser. 1993. “On Argument Structure and the Lexical Expression of Syntactic Relations”. The View from Building 20, ed. by K. Hale & S.J. Keyser, 195–240. Cambridge, Mass.: MIT Press.

Hale, K. & S. J. Keyser. 2002. Prolegomenon to a Theory of Argument Structure. Cambridge, Mass.: MIT Press.

Harley, H. 2002. “Possession and the Double Object Construction”. The Yearbook of Linguistic Variation , ed. by P. Pica and J. Rooryck, 2, 29–68.

Jackendoff, R. S. 1990. Semantic Structures. Cambridge, Mass.: MIT Press.

Jensen, F. 1990. Old French and Comparative Gallo-Romance Syntax. (Beihefte zur Zeitschrift für romanische Philologie). Tübingen: Niemeyer.

Kilroe, P. 1994. “The Grammaticalization of French à”. Perspectives on Grammaticalization, ed. by W. Pagliuca, 49–62. Amsterdam & Philadelphia: Benjamins.

Koopman, W. 1990. “The Double Object Construction in Old English”. Ms., University of Amsterdam.

Koopman, W. 1993. “The Order of Dative and Accusative Objects in Old English and Scrambling”. Studia Anglica Posnaniensia, XXV–XXVII, 109–21.

Koopman, W. 1998. “Inversion after Single and Multiple Topics in Old English”. Advances in English Historical Linguistics, ed. by J. Fisiak & M. Krygier, 135–49. Berlin: Mouton.

Krifka, M. 1999. “Manner in Dative Alternation”. Proceedings of the 18 the West Coast Conference on Formal Linguistics, ed. by S. Bird, A. Carnie, J.D. Haugen & P. Norquest, 260–271. Somerville & Medford, Mass.: Cascadilla Press.

Krifka, M. 2004. “Semantic and Pragmatic Conditions for the Dative Alternation”. Korean Journal of English Language and Linguistics 4: 1–32.

Kroch, A. & A. Taylor. 2000. “Verb-Object Order in Early Middle English”. Online posting. http://www.ling.upenn.edu/~kroch/online.html

Larson, R. K. 1988. “On the Double Object Construction”. Linguistic Inquiry 19: 335–391.

Levin, B. 1993. English Verb Classes and Alternations. Chicago: University of Chicago Press.

Levin, B. 2008. “Dative Verbs: A Crosslinguistic Perspective”. Lingvisticæ Investigationes 31: 285–312.

McFadden, T. 2002. “The Rise of the to-Dative in Middle English”. Syntactic Effects of Morphological Change, ed. by D. Lightfoot, 107–23. Oxford: Oxford University Press.

McLaughlin, J. 1983. Old English Syntax. A Handbook. Tűbingen. Niemeyer.

Mitchell, B. 1985. Old English Syntax. Oxford: Clarendon Press.

Oehrle, R. T. 1976. “The Grammatical Status of the English Dative Alternation”. Unpublished doctoral dissertation, MIT.

Pinker, S. 1989. Learnability and Cognition: The Acquisition of Argument Structure. Cambridge, Mass.: MIT Press.

Polo, C. 2002. “Double Objects and Morphological Triggers for Syntactic Case”. Syntactic Effects of Morphological Change, ed. by D. Lightfoot, 124–42. Oxford: Oxford University Press.

Rappaport Hovav, M. & B. Levin. 2005. “All Dative Verbs are not Created Equal”. Unpublished manuscript. http://www.stanford.edu/~bclevin/pubs.html

Rappaport Hovav, M. and B. Levin. 2008. “The English Dative Alternation: The Case for Verb Sensitivity”. Journal of Linguistics 44: 129–167.

Schoorlemmer, M. 1994. “Dative Subjects in Russian”. Proceedings of the Annual Workshop on Formal Approaches to Slavic Linguistics: The Ann Arbor Meeting: Functional Categories in Slavic Syntax, ed. by Jindřich Toman, 129–172. Ann Arbor, Mich.: Michigan Slavic Publications.

Siewierska, A. 1998. “Languages with and without Objects: The Functional Grammar Approach”. Languages in Contrast 1: 173–190.

Visser, F. Th. 1963–73. An Historical Syntax of the English Language, vols. I–IV. Leiden: E. J. Brill.

Weerman, F. 1997. “On the Relation between Morphological and Syntactic Case”. Parameters of Morphosyntactic Change, ed. by A. van Kemenade & N. Vincent, 427–59. Cambridge: Cambridge University Press.