The Ban on Preposition Stranding in Old English

Studia Neophilologica 77: 2–10, 2005
The Ban on Preposition Stranding in Old English
CONCHA CASTILLO
1. Introduction
As is well-known, the phenomenon of preposition stranding is highly restricted in
O(ld)E(nglish) as compared to P(resent)D(ay)E(nglish), in relation to both Amovement and A9-movement (or Wh-movement) structures: on the one hand, there
is no prepositional passive in OE (1), and on the other hand prepositions cannot
appear stranded in wh-questions (2a), relative clauses introduced by the pronoun se/
seo/ æt (the OE counterparts of PDE wh-pronouns) (2b), or topicalisation structures
introduced by a non-pronominal DP (2c). Instead, prepositions must be pied piped
along with their objects or complements. The following is, thus, a taxonomy of
structures that cannot be found in the grammar of the OE period.
(1) Peter was relied on
(2) a. Who did they rely on?
b. the man who they relied on
c. Peter, they all relied on
By contrast with (2), preposition stranding is licensed in the A9-movement structures
illustrated in (3) below, namely relative clauses introduced by the complementiser e
(3a), relatives introduced by r-pronouns (3b), topicalisation structures introduced by
a pronominal DP (3c), to-infinitive relatives (3d), and to-infinitive clauses
subcategorised by adjectives (3e).
(3) a. On am munte Synay, e se Ælmihtiga on becom,…
on the mountain Sinai which the Almighty on came
‘On mount Sinai, on which the Almighty came,…’
Ælfric’s Catholic Homilies: the Second Series 12.1.116.226
b. u ær nane myrh e on næfdest
you there no joy in had-not
‘you did not take any joy in that’
King Alfred’s Vers. of Boethius’ De consolatione philosophiae 7.15.11
c. him com æt leoht to urh paules lare
him came the light to through Paul’s teachings
‘enlightment came to him through Paul’s teachings’
King Alfred’s Vers. of St. Agustine’s Soliloquies 30.18
d. Drihten, u e gecure æt fæt on to eardienne
Lord you yourself chose that vessel in to live
‘Lord, you chose that vessel for yourself to live in’
The Blickling Homilies 157
e. ne bið ær e e in spor on to findanne
not is there easy your footstep on to find
‘your footstep is not easy to find on it’
The Paris Psalter (Krapp and Dobbie, eds.) 76.16
The aim of the present paper is to offer an analysis of why preposition stranding is
not possible in the OE counterparts of (1) and (2) above, without entering into an
DOI: 10.1080/00393270510034823
Studia Neophil 77 (2005)
The ban on preposition stranding in Old English
analysis of (3). Specifically, structures (3a, d, e) have given rise to a debate in the
literature as to whether they are the result of A9-movement (or Wh-movement), or
whether they are the product of a process of controlled unbounded deletion of a
relative pronoun (see Allen 1980 and van Kemenade 1987). As for (3b, c), the
phenomenon of preposition stranding is made possible through the use of clitic-like
elements like personal pronouns or the locative pronoun ær (see Abels 2003). The
structures in (3) are therefore involved in such processes as cliticisation or the
behaviour of covert elements, and can be dispensed with in a framework like this,
which is centred upon structures containing a PP whose object is an overt non-cliticlike nominal.
The paper is divided as follows. In section 2 I sketch the main lines of standard
GB frameworks, which are based on the notion of lexical government and reanalysis,
and which give a prominent role to inherent Case only as far as A-movement
structures or passives are concerned. In 3, I propose an approach to the ban on
preposition stranding without invoking the notion of (lexical) government and
likewise without resorting to any rule or mechanism of reanalysis. Such an approach
is based exclusively on the checking of inherent Case, which is claimed to apply
differently depending on whether it is prepositional objects or verbal objects that are
involved. Finally, a analysis of PDE preposition stranding structures is presented
which eliminates the need of reanalysis in passive configurations.
2.
The government-based approach
The ban on preposition stranding in OE is analysed in the GB literature as the failure
of OE prepositions to be proper governors – see van Kemenade (1987) or Lightfoot
(1991), modelled on works like van Riemsdijk (1978), Kayne (1981), and Hornstein
and Weinberg (1981a). The above-mentioned GB frameworks share the common
view that the traces left by the movement of the DP (or rather NP) object of P is not
properly governed by P itself, with the result that the configuration does not comply
with the Empty Category Principle or ECP. Specifically, the preposition is unable to
lexical-govern the NP trace, given that, as noted below, prepositions assign inherent
or oblique Case. The OE situation is solved in M(iddle)E(nglish) by a reanalysis rule
that turns V and P into a kind of compound or complex verb that is actually able to
govern what was originally the object or complement of the preposition (see (5)
below).
(4) *[DPi [VPV [PPP ti]]]
(5) [DPi [VP[VV P] [DPti]]]
Although there seems to be agreement that the reanalysis process in A9-movement
structures like interrogatives is not so strict or severe as the reanalysis process
applying in A-movement configurations or passives, given that only the latter seems
to require adjacency between V and P,1 in both A- and A9-movement structures P
must ultimately become tightly connected with V according to GB approaches.
Thus, van Kemenade (1987: 216ff.), following Hornstein and Weinberg (1981b),
argues in favour of a mechanism or process of index transmission instead of
reanalysis proper in the case of A9-movement constructions.
Though the idea of the failure of government and the subsequent reanalysis
process characterises both A-movement and A9-movement prepositional configurations from a diachronic perspective, the above-mentioned GB literature agrees that
stranding by passivisation, that is movement of a prepositional object in a
A-movement structure like the one illustrated in (1) above, further requires for the
3
4
C. Castillo
Studia Neophil 77 (2005)
Case assigned by P to be of the structural type, and not of the inherent or oblique
type. Prepositions in OE have generally been assumed to assign Case at D-structure
– the so-called inherent or oblique Case. Following the GB literature, for prepositions
to assign inherent Case affects the lack of preposition stranding in OE A-movement
configurations but not in A9- or Wh-movement structures: as mentioned above,
A9-movement structures are exclusively affected by the change occurring in ME
relative to the availabiblity of a reanalysis mechanism (or index transmission
process). By contrast with A9-movement structures, the movement that applies in
passives or A-movement implies that the moved constituent is assigned Case
– specifically, nominative Case – in the target position. Thus, for a prepositional
passive configuration to be instantiated it is necessary not only that the abovementioned reanalysis process takes place but also that the preposition assigns no
inherent Case to its object or complement prior to the movement of the latter: if,
contrary to this, the preposition gives inherent Case to its complement, there will be
a Case clash between such inherent Case and the nominative Case that the NP object
is assigned at S-structure in subject position.
This way, the frameworks or analyses following the GB model of grammar defend
the idea that preposition stranding by both passivisation and Wh-movement is
prohibited in OE because there is no reanalysis rule that can solve the problem of the
violation of ECP. Apart from prepositions becoming proper governors, prepositional passives demand that prepositions not assign inherent Case to their objects or
complements. Let us note at this point that though verbs can also assign inherent
Case in OE, they do not fail to govern the traces left by their objects, which means
that the latter can actually move – see the illustrations in 3.1 below. By contrast with
prepositions, verbs assigning inherent Case are therefore proper (lexical) governors.
The crucial observation to be made now is that the concept of government does
not belong within the apparatus of minimalism. The framework that is proposed
below aims at explaining the ban on preposition stranding in OE without resorting
to lexical government as exercised by prepositions.
3. A characterisation of inherent Case
I would like to propose that the ban on preposition stranding in OE can be explained
by appealing to the inherent Case checked by prepositions in both A-movement and
A9-movement configurations. As noted above, inherent Case is a determining factor
in former GB analyses of stranding structures, but only as far as passive
configurations are concerned. In the framework proposed here, inherent Case is
responsible for the indivisible nature of Prepositional Phrases in OE, whether these
occur in A-movement or in A9-movement environments.
According to the GB tradition (Chomsky 1981, 1986), inherent Case is
distinguished from structural Case in that it is assigned at D-structure and is
closely connected with thematic properties, which means that it is assigned
through the mechanism of lexical government. By contrast, structural Case
presupposes head government at S-structure and is dissociated from any thematic
properties. Turning now to diachronic facts, it must be noted that whereas both
verbs and prepositions assign structural Case in PDE, OE verbs can assign either
structural Case or inherent Case – specifically genitive, as in (6a) or dative, as in (6b)2
– and all prepositions generally speaking assign inherent Case – specifically dative,
or (oblique) accusative (see (7) below).3,4 As a matter of fact, the loss of inherent
or oblique Case at the end of the OE period and the beginning of the ME period is
one of the most decisive morphosyntactic changes affecting the grammar of the
language.
Studia Neophil 77 (2005)
The ban on preposition stranding in Old English
(6) a. Se Themestocles gemyndgade Ionas ære ealdan fæh egenit.
this Themestocles reminded Ionians the old vendetta
‘This Themestocles reminded the Ionians of the old vendetta’
The Old English Orosius 47,20
b. Helpa earmumgenit. 7 hæfenleasumgenit.
help poor and needy
‘Help the poor and the needy’
The Homilies of Wulfstan 11.197
(7) a. he wolde mid his freondumdat gesprec habban
he wanted with his friends talk have
‘he wanted to have talk with his friends’
The OE Vers.of Bede’s Ecclesiastical History of the English People 134.3
b. ðone Halgan Gast, ðurh ðoneaccus. we beoð gehalgode5
‘the Holy Ghost, through whom we are hallowed’
Ælfric’s Catholic Homilies: the Second Series 17.167.190
As mentioned above, inherent Case is primarily characterised in the GB model
as applying at D-structure (as opposed to structural Case, which takes place at
S-structure). However, no such levels exist any longer in minimalist theories of
grammar. According to current minimalist accounts (Chomsky 1995, 2001), the
assignment – or rather checking – of Case takes place in Spec-head agreement
relations. Movement or attraction is assumed to be driven by the requirement of
feature-checking, Case being one such feature. A relation of agreement is established
between the Spec position of a functional projection and the head of the latter, and
the constituent in need of checking its features must move to such Spec position.
This way, nominative Case is checked or licensed in the Spec position of TP in
English, and accusative Case is checked in the Spec position of an AgrOP projection
within vP.6
Such Spec-head agreement relations constitute the environment where structural Case takes place. As far as inherent Case is concerned, there seems to be no
detailed characterisation of its properties apart from the assumption that it is
checked in situ. That is, whereas structural Case implies movement of the DP in
need of checking its Case-features to the Spec position of a functional projection,
inherent Case takes place between the lexical head itself and its DP object or
complement without the assistance of any functional projection. The Case of the
verbal objects in structures like the ones in (6) above would then be checked or
licensed by such objects occupying the position sister to V; similarly, the Case of
the prepositional objects in (7) is checked by the latter in the position that is sister
to P.
According to minimalist theories, a crucial property that is at the base of the
mechanism of Case-checking is whether such a mechanism applies before Spell-Out or
otherwise after Spell-Out, at the level of LF. Movement applying before Spell-Out is
overt movement and can therefore be appreciated in the phonetic chain. By contrast,
movement taking place at LF is covert movement, and is invisible from the point of
view of phonetic realisation. Overt movement is assumed to be paired with the
checking of strong features, whereas covert movement is coupled with weak features.
The above-mentioned distinction between overt and covert movement is generally
applied to structural Case, that is the Case that is licensed through Spec-head
agreement configurations. In the framework proposed here, a parallel differentiation
is assumed to affect inherent Case. The ban on preposition stranding in OE – and, as
a matter of fact, in a vast number of (contemporary) languages – is viewed here as
the result of prepositions checking against their DP objects or complements a type
of inherent Case that does not allow for the division or decomposition of the
Prepositional Phrase itself.
5
6
C. Castillo
3.1.
Studia Neophil 77 (2005)
The proposed framework: inherent Case in OE
As observed in the preceding sections, OE prepositions check inherent Case, in a
similar fashion to a vast number of verbs. However, whereas the DP object of
V checking inherent Case can detach itself from V in both A’-movement and
A-movement configurations, this is not possible for the DP object of P, hence the
ban on preposition stranding.
(8a) below illustrates a topicalisation structure where an object of V, which
happens to check genitive, has been raised to [Spec, CP].7 As for (8b), it is an
example of so-called impersonal passives, a configuration where a passive verb
subcategorises for an inherent object marked with dative Case.
(8) a. And æsgenit. us ne scamað na
and of that to-us not shames never
‘And we are not ashamed of that at all’
The Homilies of Wulfstan 20.3 160
b. Ac ðæmdat. mæg beon suiðe hraðe geholpen
but to-him may be very quickly helped
‘But he may be helped very quickly’
King Alfred’s West Saxon Vers. of Gregory’s Pastoral Care 33.225.22
Instead of claiming that verbs, but not prepositions, are proper governors in
OE (as in standard GB frameworks), I propose that the type of inherent Case
checked or licensed by verbs is of a different nature than the type of inherent
Case checked by prepositions. Specifically, I would like to make the point that
verbal objects check their inherent Case in an overt manner. This way, verbal objects
check their Case in situ before Spell-Out, and are free to move afterwards to any
other position within clausal structure: in (8a) above, the DP object of V has moved
into [Spec, CP]; as for (8b), I will not enter here into the discussion of whether the
dative DP is in a topicalised position or in the position of subject: what matters
from the perspective of the present approach is that the DP has actually moved
away from V.
By contrast with oblique objects of verbs, prepositional objects cannot be
extracted from P in order to be questioned, relativised, topicalised, or passivised.
Such distinct behaviour can be explained by claiming that prepositional objects
check their Case in situ (inherent or oblique Case) but in a covert manner, that is
after Spell-Out. According to the Principle of Full Interpretation, if the DP object of
P does not have its Case-features checked before Spell-Out, the DP itself cannot
count as a linguistic element when moving overtly to a higher position within CP, or
to the subject position. Thus, though oblique objects or complements of both V and
P check their Case in situ, it is radically different for such Case-checking to apply
before Spell-Out, or otherwise at LF: only those objects checking their inherent Case
overtly will be visible to further overt movements.
The ban on preposition stranding in OE, that is the inability of prepositional
objects to be extracted from Prepositional Phrases, through either A-movement or
A9-movement, seems to be intimately connected with a word order property in
ordinary, non-stranding structures which again opposes Prepositional Phrases to
Verb Phrases: as is well-known, (non-pronominal) objects of P invariably follow P,
whereas (non-pronominal) objects of V can appear either to the left or to the right of
V, hence the SOV and SVO orders, respectively. Whether the basic word order in OE
is SOV or SVO is not relevant for the present discussion (the reader is referred to
Fischer et al. (2000) for an introductory study): what matters is that, by contrast with
VPs, word order inside PPs is a fixed one, with no possibility for the object of P to
occupy any other position.
Studia Neophil 77 (2005)
The ban on preposition stranding in Old English
(9) a. [PPP O]
b. [VPV O]/[VPO V]
Positing that prepositions and their objects check inherent Case against each other in
a covert manner, as opposed to verbs and their objects, eliminates the need to resort
to government as an explanation of the lack of preposition stranding structures in
OE, both in A-movement and in A9-movement configurations. Recall from section 2
above that GB theory used to argue that the failure of (lexical) government was
solved by means of a reanalysis mechanism between V and P entering the grammar
of English in the ME period. In a framework like the one presented here, which does
not invoke the concept or mechanism of government, no reanalysis process is
necessary either. It must be noted at this point that reanalysis seems generally
speaking not to be a desirable mechanism from the point of view of minimalism (see
Baltin and Postal 1996).
In 3.2 below I will deal with the analysis of preposition stranding structures in
PDE, which is actually available immediately after the OE period. It will be observed
that the possibility seems to exist for the analysis of PDE prepositional passives to
dispense with the mechanism of reanalysis.
Before putting an end to this section, I would like to note that the grammar of
Spanish or French, or indeed that of a vast number of well-known languages lacking
preposition stranding configurations, is to be analysed in a parallel fashion to OE
grammar. As the following illustrations show, objects cannot be extracted from PPs
in Spanish, which means (i) that the only possibility in A9-movement structures like
(10) is the mechanism of pied piping, and (ii) that no prepositional passive exists in
the language (see the ungrammaticality of (11) below). In a parallel fashion to OE
prepositional objects, Spanish prepositional objects would check covert inherent
Case.
(10) a. ¿Con quién trabaja Juan?/El anticuario para quien trabaja Juan…
with whom works Juan the antiquarian for whom works Juan
‘With whom does Juan work?’/‘The antiquarian for whom J. works…’
b. *¿Quién trabaja Juan para?/*El anticuario que Juan trabaja para…
who works Juan for the antiquarian that Juan works for
(11) *Marı́a fue reı́da de (por todos)
Marı́a was lauged of by everybody
‘Marı́a was laughed at (by everybody)’
3.2.
Preposition stranding in PDE
Preposition stranding configurations begin to appear in ME: specifically
A9-movement structures start to be available at the beginning of the thirteenth
century, and the first prepositional passives are found in the first half of the
thirteenth century. Once prepositional objects begin checking structural Case,
English PPs stop being indivisible constituents. In a similar fashion to verbal objects,
prepositional objects are now able to move away from the head of their own
projections, a situation that lasts into PDE.
As noted in the preceding section, preposition stranding is banned in Spanish,
among many other languages, which means that prepositions check inherent Case in
these languages, specifically covert inherent Case. By contrast with both English and
Spanish, Icelandic represents the type of language where extraction from PP is
permitted, though prepositional objects check inherent Case. Following the present
framework, prepositions in Icelandic check overt inherent Case.
7
8
C. Castillo
Studia Neophil 77 (2005)
(12) a. Hann spur2i hvern eg hef2i tala2 vi2
he asked whomaccus I had talked with
b. ennan ref hefur aldrei veri skoti2 á
that foxaccus has never been shot at
(van Kemenade 1987: 217)
Though the main purpose of this paper was to offer an explanation of the ban on
preposition stranding in OE without invoking the notion of government, and
additionally that of reanalysis, I would like to offer a characterisation of preposition
stranding in PDE in both passives or A-movement structures and Wh- or A9movement configurations.
As argued above, preposition stranding becomes available in the English language
once prepositional objects stop checking covert inherent Case against their prepositions. From a minimalist perspective, A9-movement structures like the interrogatives
in (13) below are characterised by the attraction of a constituent carrying a whfeature into [Spec, CP], an approach that is already advanced in the GB model of
grammar (see Chomsky 1981). This way, the constituent moving into [Spec, CP] in
the stranding structure (13a) is an interrogative DP, whereas the one in the pied
piping configuration (13b) is a full PP. This means that the preposition, though not a
wh-item itself, is marked with a [+wh]-feature in pied piping structures. Let us note
that, according to the present approach, the reason why a language like Spanish does
not admit (14a) – and similarly (10b) above – is not that Spanish prepositions are
obligatorily marked [+wh], but that Spanish prepositions check covert inherent Case
and therefore their objects cannot move away from them.
(13) a. Who did John talk with?
b. With whom did John talk?
(14) a. *¿Quién habló Juan con?
who talked Juan with
b. ¿Con quién habló Juan?
with whom talked Juan
The analysis presented above clearly shows that the concept of government has no
place in PDE stranding by Wh-movement from a minimalist perspective. Similarly,
no reanalysis rule or mechanism is invoked between V and P: for a prepositional
object to be wh-moved does not mean that P is reanalysed with V and that therefore
the object of P becomes the object of [VV P], but that the only constituent that is
marked [+wh] is the wh-item itself, and not P.
However, in spite of the fact that reanalysis is not a much wanted solution from
the point of view of minimalism, the current analysis of prepositional passives is
still dependent upon such a mechanism. Thus, it is generally assumed that the
ungrammaticality of a sequence like (15) below is due to the Case of the DP John not
being checked. Though John is originally the object or complement of P, a renalysis
process by means of which P is incorporated into V turns John into the object of
[VV P], and the latter is eventually affected by the inability of passive V to check
(accusative) Case.
(15) *It was looked at John
In the remainder of this section I will advance an analysis of prepositional passives
which does not invoke any reanalysis rule. I must note that this topic is currently
under research by the author.
Studia Neophil 77 (2005)
The ban on preposition stranding in Old English
As is well-known, DP objects or complements of V are currently assumed to be
generated as sisters of V, and to raise from that position into a functional projection
in order to check their features. Following Bobaljik (1995), Koizumi (1995), or
Bowers (2002), I assume that the above-mentioned functional projection, which is
called TrP in Bowers (2002),8 is located within the vP configuration, specifically
above VP. As shown in (16) below, the verbal object raises into the Spec of TrP.
(16) a. [vP [TrP [VP V DP]]]
b. [vPVv [TrP DPi[VP tv ti]]]
In a similar fashion to DP complements of V, PP complements of V can be argued to be
related to V through a functional projection, PrepTrP, located in between vP and VP.
(17) a. [vP [PrepTrP [VP V PP]]]
b. [vPVv [PrepTrPPPi [VP tv ti]]]
A crucial difference between DP complements and PP complements of V is that only
the former can passivise, which is precisely why Bowers (2002) defends the presence
of TrP also in passive structures (let us note that the author deals only with ordinary
or non-prepositional passives). I would like to argue that a passive structure like (18)
(former (15)), where V is followed b a PP complement lacks any PrepTrP, with the
result that the PP cannot check its features, and the derivation crashes. That PPs are
not passivisable objects is clearly shown in (19).
(18) *It was looked at John
(19) *At John was looked
Though the full PP is not licensed as object or complement of V, its DP object is.
Thus, English allows for the possibility that the prepositional object raises to [Spec,
TP] in order to check its own features, thereby leaving the preposition stranded. In a
framework like this, there is no need for a reanalysis rule or mechanism between V
and P that turns the prepositional object into a verbal object (see (5) above): instead,
the PP is simply not licensed as such phrasal projection, but the object of P is, and as
a result P is left stranded. This means that reanalysis is neither the cause nor the
explanation of the phenomenon of preposition stranding.
4.
Conclusion
It has been argued that, by applying to inherent Case the well-known distinction or
differentiation between overt vs. covert movement, the ban on preposition stranding
in OE (and in a vast number of contemporary languages) can be explained without
resorting to lexical government and additionally to reanalysis. The inability of OE
prepositional objects to be extracted from PP in both A-movement and A9movement structures has been argued to be due to Case-checking occurring after LF
and to the subsequent violation of the Principle of Full Interpretation. PDE
stranding by passivisation has likewise been explained without invoking any
reanalysis rule, which would seem to be in accord with a minimalist perspective.
Dpt. of English
University of Málaga (Spain)
Facultad de Filosofı́a y Letras
Campus de Teatinos
29071 MÁLAGA
9
10 C. Castillo
Studia Neophil 77 (2005)
NOTES
1 Note the contrast between the PDE structures below: (i) is an interrogative sentence, whereas (ii)
illustrates a passive configuration.
(i.) Whose mother did John shout angrily at?
(ii.) *Beth’s mother was shouted angrily at
2 As is well-known, OE adjectives, in a similar fashion to verbs, can assign (inherent) dative or genitive
Case.
3 The inherent accusative Case assigned or checked by prepositions is not distinguished morphologically
speaking from the structural accusative Case checked by verbs, it being precisely the inability of the
prepositional object to be extracted from P that characterises the former.
4 Prepositions very rarely assign or check genitive Case.
5 In this very case, the preposition assigns inherent accusative to a relative pronoun in the complementiser position.
6 See, however, section 3.2 below for another name for the same projection.
7 Additionally, the verb is an impersonal one taking a dative DP.
8 In frameworks like those of Bobaljik (1995) or Koizumi (1995), such a functional projection bears the
name of AgrOP, which is reminiscent of the phrase occurring in between TP and VP previous to the
advent of the vP-shell mechanism.
REFERENCES
Abels, Klaus (2003). ‘‘A note on clitics and P-stranding’’. Formal Approaches to Slavic Linguistics 11.
Allen, Cynthia (1980). ‘‘Movement and deletion in Old English’’. Linguistic Inquiry 11: 261–323.
Baltin, Mark & Paul M. Postal (1996). ‘‘More on reanalysis hypotheses’’. Linguistic Inquiry 27: 127–145.
Bethurum, D. (ed.) (1957). The Homilies of Wulfstan. Oxford, Clarendon Press.
Bobaljik, Jonathan D. (1995). ‘‘Morphosyntax: the syntax of verbal inflection’’. MIT Ph.D. dissertation.
Bately, J. (ed.) (1980). The Old English Orosius. EETS SS6.
Bowers, John (2002). ‘‘Transitivity’’. Linguistic Inquiry 33: 183–224.
Carnicelli, T. A. (ed.) (1969). King Alfred’s Version of St. Augustine’s Soliloquies. Cambridge, Mass.,
Harvard U.P.
Chomsky, Noam (1981). Lectures on Government and Binding. The Pisa Lectures. Dordrecth, Foris.
Chomsky, Noam (1986). Knowledge of language: its nature, origin, and use. New York, Praeger.
Chomsky, Noam (1995). The Minimalist Program. Cambridge, Mass., MIT Press.
Chomsky, Noam (2001). ‘‘Beyond explanatory adequacy’’. Ms. Cambridge, Mass., MIT.
Fischer, Olga et al. (2000). The syntax of Early English. Cambridge, Cambridge U.P.
Godden, M. (ed.) (1979). ÆLfric’s Catholic Homilies: the Second Series. EETS SS 5.
Hornstein, Norbert & Amy Weinberg (1981a). ‘‘Case theory and preposition stranding’’. Linguistic
Inquiry 12: 55–93.
Hornstein, Norbert & Amy Weiberg (1981b). ‘‘Extraction, government and parameters of grammar’’.
Unpublished ms.
Kayne, Richard (1981). ‘‘ECP extensions’’. Linguistic Inquiry 12: 93–135.
Kemenade, Ans van (1987). Syntactic case and morphological case in the history of English. Dordrecht,
Foris.
Koizumi, Masatashi (1995). ‘‘Phrase structure in minimalist syntax’’. MIT Ph.D. dissertation.
Krapp, G. P & E. V. K. Dobbie (eds.) (1931–1953). The Anglo-Saxon Poetic Records: a Collective Edition.
New York and London, Columbia U.P.
Lightfoot, David (1991). How to set parameters: arguments from language change. Cambridge, Mass., MIT
Press.
Miller, Th. (ed.) (1890–1898). The Old English Version of Bede’s Eclesiastical History of the English People.
EETS 95, 96, 110, 111.
Morris, R. (ed.) (1880). The Blickling Homilies of the tenth century. EETS 58, 63,73, Oxford, N. Trubner.
Riemsdijk, Henk van (1978). A case study in syntactic markedness: the binding nature of Prepositional
Phrases. Dordrecht, Foris.
Sedgefield, W. J. (ed.) (1899). King Alfred’s Old English Version of Boethius’s De consolatione philosophia.
Oxford, Clarendon Press.
Sweet, H. (ed.) (1871). King Alfred’s West Saxon Version of Gregory’s PastoralCare. EETS 45, 50.