From GB to minimalism Marcel den Dikken ELTE

From GB to minimalism Marcel den Dikken ELTE & RIL-HAS Budapest

Marcel den Dikken @ EGG 2016 Government & Binding Theory Chomsky (1981), Lectures on Government and Binding, revolutionised generativesyntactic theorising, by directing the field away from a system of rules (typically language- and/or construction-specific) and (typically ad hoc) filters on rules and towards a system of universal principles and parameters capturing variation 2

Marcel den Dikken @ EGG 2016 Government & Binding Theory though widely referred to at the time as ‘Government & Binding Theory’ (also by its inventor himself: see e. g. Chomsky 1982), Chomsky has come to prefer the designation ‘Principles & Parameters Theory’ (P&P): ‘government’ and ‘binding’ are important ingredients of theory but they are not in any way their essence 3

Marcel den Dikken @ EGG 2016 4 Government & Binding Theory Chomsky (1995: 29– 30): ‘The P&P approach is sometimes termed Government-Binding (GB) Theory. The terminology is misleading. True, early efforts to synthesize current thinking in these terms happened to concentrate on theories of government and of binding. . . , but these modules of language stand alongside many others: Case theory, θ-theory, and so on. It may turn out that the concept of government has a kind of unifying role, but there is nothing inherent to the approach that requires this. ’

Marcel den Dikken @ EGG 2016 Government & Binding Theory the essence of GB/P&P theory: • principles • parameters • modules 5

Marcel den Dikken @ EGG 2016 6 A Modular Theory P&P theory is modular in that (a) it features different levels of representation (D-structure, S-structure, LF, PF) (b) it features multiple subtheories (Case theory, θ-theory, trace theory, bounding theory, binding theory, etc. )

Marcel den Dikken @ EGG 2016 A Modular Theory to determine whether a particular syntactic construct is grammatical or ungrammatical, theory consults all the relevant modules that bear on the construct, and determines for each of them whether or not any of its principles and parameters is violated to any construct that violates something in one or more modules the grammar gives a * 7

Marcel den Dikken @ EGG 2016 A Modular Theory a construct that violates multiple principles is expected to be more severely penalised than one that violates just one there has never been a principled theory about the possibility that a violation of a principle in module A could be more costly than a violation of a principle in module B 8

Marcel den Dikken @ EGG 2016 Subjacency & ECP it has been standard practice, however, to treat a violation of the Subjacency Condition (the core principle of bounding theory) as less severe than a violation of, say, the Empty Category Principle (at the heart of trace theory) it is often said that a Subjacency violation ‘feels less bad’ than an ECP violation 9

Marcel den Dikken @ EGG 2016 Subjacency & ECP Subjacency Condition movement must not traverse more than one bounding node (English: IP, NP) at a time Empty Category Principle a trace must be properly governed (a) head government, or (b) antecedent government 10

Marcel den Dikken @ EGG 2016 11 Subjacency & ECP Subjacency Condition movement must not traverse more than one bounding node (English: IP, NP) at a time ugly because it counts (‘more than one’), and stipulates bounding nodes — and does so in different ways for different languages

Marcel den Dikken @ EGG 2016 12 Subjacency & ECP ugly because it is disjunctive and because by their nature, ‘head government’ and ‘antecedent government’ are two very diverse notions (i. e. , the ECP is a fake unification) Empty Category Principle a trace must be properly governed (a) head government, or (b) antecedent government

Marcel den Dikken @ EGG 2016 13 Barriers — The Mission recognising that LGB’s approach to the licensing and bounding conditions on movement (‘Move α’) is far from optimal, Chomsky (1986, Barriers) seeks to both unify and simplify theories of trace licensing (ECP) and bounding (Subjacency) and to eliminate the stipulativeness of the latter by replacing it with an algorithmic approach to the computation of barriers

Marcel den Dikken @ EGG 2016 14 Barriers — The Mission Subjacency Condition movement must not traverse a barrier NB 1: Barriers uses ‘ 0 -subjacency’ esp. in the analysis of parasitic gap constructions; but elsewhere it works perfectly fine, too Empty Category Principle a trace must be antecedent governed NB 2: Barriers keeps the LGB-style disjunctive ECP; but it can actually do without head gov’t

Marcel den Dikken @ EGG 2016 Barriers — The Mission Subjacency Condition movement must not traverse a barrier Empty Category Principle a trace must be antecedent governed compare this to what we had in LGB… 15

Marcel den Dikken @ EGG 2016 Flashback… Subjacency Condition movement must not traverse more than one bounding node (English: IP, NP) at a time Empty Category Principle a trace must be properly governed (a) head government, or (b) antecedent government 16

Marcel den Dikken @ EGG 2016 Barriers — The Mission Subjacency Condition movement must not traverse a barrier Q 1: how do we identify a ‘barrier’? Empty Category Principle a trace must be antecedent governed Q 2: how do we define ‘government’? 17

Marcel den Dikken @ EGG 2016 18 Barriers — The Mechanics government α governs β iff (i) and (ii) (i) α c-commands β (ii) no barrier intervenes between α and β Q 2: how do we define ‘government’?

Marcel den Dikken @ EGG 2016 19 Barriers — The Mechanics barrier a domain Δ is a barrier for a relation R between α and β iff Δ dominates β and excludes α and (i) or (iii) (i) Δ is a blocking category for β (Δ ≠ IP) (ii) Δ is a maximal projection immediately dominating a blocking category for β (iii) Δ is the immediate projection of γ, a governor of β

Marcel den Dikken @ EGG 2016 20 Barriers — The Mechanics barrier a domain Δ is a barrier for a relation R between α and β iff Δ dominates β and excludes α and (i) or (iii) (i) Δ is a blocking category for β (Δ ≠ IP) (ii) Δ is a maximal projection immediately dominating a blocking category for β (iii) Δ is the immediate projection of γ, a governor of β

Marcel den Dikken @ EGG 2016 21 Barriers — The Mechanics barrier a domain Δ is a barrier for a relation R between α and β iff Δ dominates β and excludes α and (i) or (iii) (i) Δ is a blocking category for β (Δ ≠ IP) (ii) Δ is a maximal projection immediately dominating a blocking category for β Q 3: how do we identify a ‘blocking category’?

Marcel den Dikken @ EGG 2016 22 Barriers — The Mechanics blocking category a projection Π is a blocking category for β iff Π dominates β and is not θ-governed by a lexical category (‘L-marked’) by clause (i), a blocking category for β is almost always a barrier for β as well — but an exception is made for IP

Marcel den Dikken @ EGG 2016 23 Barriers — The Mechanics barrier a domain Δ is a barrier for a relation R between α and β iff Δ dominates β and excludes α and (i) or (iii) (i) Δ is a blocking category for β (Δ ≠ IP) if IP were automatically a barrier for β by being a blocking category for β we would threaten to exclude all syntactic relations across an IP in a non-θ-governed position

Marcel den Dikken @ EGG 2016 24 Barriers — The Mechanics barrier a domain Δ is a barrier for a relation R between α and β iff Δ dominates β and excludes α and (i) or (iii) (i) Δ is a blocking category for β (Δ ≠ IP) since sentences of the following type are grammatical, a proper antecedent government relation must be establishable across IP [CP how [C=did [IP she [VP do it t ]]]]?

Marcel den Dikken @ EGG 2016 25 Barriers — The Mechanics barrier a domain Δ is a barrier for a relation R between α and β iff Δ dominates β and excludes α and (i) or (iii) (i) Δ is a blocking category for β (Δ ≠ IP) in this example, we don’t need to worry about VP as a potential barrier: t is adjoined to VP, hence not dominated by VP, so VP ≠ BC for t [CP how [C=did [IP she [VP do it t ]]]]?

Marcel den Dikken @ EGG 2016 26 Barriers — The Mechanics barrier a domain Δ is a barrier for a relation R between α and β iff Δ dominates β and excludes α and (i) or (iii) (i) Δ is a blocking category for β (Δ ≠ IP) and because VP ≠ BC for t, IP cannot inherit barrierhood for t from VP, via clause (ii), either [CP how [C=did [IP she [VP do it t ]]]]?

Marcel den Dikken @ EGG 2016 27 Barriers — The Mechanics barrier a domain Δ is a barrier for a relation R between α and β iff Δ dominates β and excludes α and (i) or (iii) (i) Δ is a blocking category for β (Δ ≠ IP) (ii) Δ is a maximal projection immediately dominating a blocking category for β → so we’re good for local adjunct extraction [CP how [C=did [IP she [VP do it t ]]]]?

Marcel den Dikken @ EGG 2016 28 Barriers — The Mechanics barrier a domain Δ is a barrier for a relation R between α and β iff Δ dominates β and excludes α and (i) or (iii) (i) Δ is a blocking category for β (Δ ≠ IP) (ii) Δ is a maximal projection immediately dominating a blocking category for β but now what about sentences such as: [CP who [C=did [IP she [VP find [AP t nice ]]]]]?

Marcel den Dikken @ EGG 2016 29 Barriers — The Mechanics barrier a domain Δ is a barrier for a relation R between α and β iff Δ dominates β and excludes α and (i) or (iii) (i) Δ is a blocking category for β (Δ ≠ IP) (ii) Δ is a maximal projection immediately dominating a blocking category for β how can t be properly governed? ? ? [CP who [C=did [IP she [VP find [AP t nice ]]]]]?

Marcel den Dikken @ EGG 2016 30 Barriers — The Mechanics barrier a domain Δ is a barrier for a relation R between α and β iff Δ dominates β and excludes α and (i) or (iii) (i) Δ is a blocking category for β (Δ ≠ IP) (ii) Δ is a maximal projection immediately dominating a blocking category for β Q 1: VP = BC and inherent barrier for t? YES! [CP who [C=did [IP she [VP find [AP t nice ]]]]]?

Marcel den Dikken @ EGG 2016 31 Barriers — The Mechanics barrier a domain Δ is a barrier for a relation R between α and β iff Δ dominates β and excludes α and (i) or (iii) (i) Δ is a blocking category for β (Δ ≠ IP) (ii) Δ is a maximal projection immediately dominating a blocking category for β Q 2: does IP inherit barrierhood for t? YES! [CP who [C=did [IP she [VP find [AP t nice ]]]]]?

Marcel den Dikken @ EGG 2016 32 Barriers — The Mechanics as analysed, this sentence should be woeful … … but it’s perfectly acceptable … … so something isn’t quite right Chomsky’s solution: intermediate adjunction to VP [CP who [C=did [IP she [VP find [AP t nice ]]]]]?

Marcel den Dikken @ EGG 2016 33 Barriers — The Mechanics as analysed, this sentence should be woeful … … but it’s perfectly acceptable … … so something isn’t quite right Chomsky’s solution: intermediate adjunction to VP [CP who [C=did [IP she [VP t′ [VP find [AP t nice ]]]]]]?

Marcel den Dikken @ EGG 2016 34 Barriers — The Mechanics barrier a domain Δ is a barrier for a relation R between α and β iff Δ dominates β and excludes α and (i) or (iii) (i) Δ is a blocking category for β (Δ ≠ IP) (ii) Δ is a maximal projection immediately dominating a blocking category for β now VP ≠ BC for either R(t, t′) or R(t′, who) [CP who [C=did [IP she [VP t′ [VP find [AP t nice ]]]]]]?

Marcel den Dikken @ EGG 2016 35 Barriers — The Mechanics barrier a domain Δ is a barrier for a relation R between α and β iff Δ dominates β and excludes α and (i) or (iii) (i) Δ is a blocking category for β (Δ ≠ IP) (ii) Δ is a maximal projection immediately dominating a blocking category for β … hence VP ≠ barrier for R(t, t′) or R(t′, who) [CP who [C=did [IP she [VP t′ [VP find [AP t nice ]]]]]]?

Marcel den Dikken @ EGG 2016 36 Barriers — The Mechanics barrier a domain Δ is a barrier for a relation R between α and β iff Δ dominates β and excludes α and (i) or (iii) (i) Δ is a blocking category for β (Δ ≠ IP) (ii) Δ is a maximal projection immediately dominating a blocking category for β and IP cannot inherit barrierhood for R(t′, who) [CP who [C=did [IP she [VP t′ [VP find [AP t nice ]]]]]]?

Marcel den Dikken @ EGG 2016 37 Barriers — The Mechanics Chomsky doesn’t provide empirical arguments for intermediate adjunction to VP in Barriers there is a theoretical advantage: all legitimate traces are now antecedent governed Chomsky’s solution: intermediate adjunction to VP even for cases of mov’t of an internal argument (θ-governed by V), there is a local antecedent that can license the original trace for the ECP [CP who [C=did [IP she [VP t′ [VP kiss t ]]]]]?

Marcel den Dikken @ EGG 2016 38 Barriers — The Mechanics Chomsky doesn’t provide empirical arguments for intermediate adjunction to VP in Barriers he also uses it in his analysis of parasitic gap constructions (truly the linchpin of Barriers) Chomsky’s solution: intermediate adjunction to VP it is interesting that in current minimalist syntax intermediate adjunction to v. P again is the norm — and again is defended with parasitic gaps [CP who [C=did [IP she [VP t′ [VP kiss t ]]]]]?

Marcel den Dikken @ EGG 2016 39 Barriers — The Mechanics Chomsky’s solution: intermediate adjunction to VP BUT there is what appears to be a simpler solution to the ‘VP as BC and inherent barrier’ problem [CP who [C=did [IP she [VP t′ [VP kiss t ]]]]]?

Marcel den Dikken @ EGG 2016 40 Barriers — The Mechanics blocking category a projection Π is a blocking category for β iff Π dominates β and is not θ-governed by a lexical category (‘L-marked’) if we replace ‘θ-governed by a lexical category’ with ‘selected’ (Chomsky & Lasnik 1993), and recognise that I selects VP, a VP in the complement of I no longer qualifies as a BC

Marcel den Dikken @ EGG 2016 41 Barriers — The Mechanics blocking category a projection Π is a blocking category for β iff Π dominates β and is not selected if we replace ‘θ-governed by a lexical category’ with ‘selected’ (Chomsky & Lasnik 1993), and recognise that I selects VP, a VP in the complement of I no longer qualifies as a BC

Marcel den Dikken @ EGG 2016 42 Barriers — The Mechanics as it happens, Chomsky (1986: 20) himself presents an argument for considering the relation between I and its VP-complement to be one of θ-government he points to the relative acceptability of the VP -fronting example below as an indication that I θ-marks its complement ? (? )fix the car, I wonder whether he will t

Marcel den Dikken @ EGG 2016 43 Barriers — The Mechanics θ-government per se does not exempt the VP in the complement of I from being a BC given the definition of BC in Barriers — but it does if we adopt the definition in terms of selection θ-government of VP by I compels us to give up intermediate adjunction to VP if we accept Chomsky’s (1986: 6) condition on adjunction: adjunction is possible only to a maximal projection that is a non-argument ? (? )fix the car, I wonder whether he will t

Marcel den Dikken @ EGG 2016 44 Barriers — The Mechanics on the standard assumption that something that is θ-governed is an argument (more generally, anything that receives a θ-role is an argument), the VP in the complement of I, which Chomsky says is θ-governed by I, is an argument, hence by the constraint below not a legitimate intermediate adjunction site adjunction is possible only to a maximal projection that is a non-argument ? (? )fix the car, I wonder whether he will t

Marcel den Dikken @ EGG 2016 45 Barriers — The Mechanics Chomsky provides a plausible rationale for his condition on adjunction if this condition stands, and VP-fronting shows that the complement of I is an argument, then this also rules out the intermediate adjunction to v. P popular in current minimalism adjunction is possible only to a maximal projection that is a non-argument ? (? )fix the car, I wonder whether he will t

Marcel den Dikken @ EGG 2016 46 Barriers — The Mechanics back to Who did she find nice? … if “Chomsky’s solution” targets a non-problem as far as VP’s inherent barrierhood and IP’s barrierhood by inheritance are concerned … Chomsky’s solution: intermediate adjunction to VP … don’t we still have to reckon with clause (iii) of the definition of a barrier? [CP who [C=did [IP she [VP find [AP t nice ]]]]]?

Marcel den Dikken @ EGG 2016 47 Barriers — The Mechanics barrier a domain Δ is a barrier for a relation R between α and β iff Δ dominates β and excludes α and (i) or (iii) (i) Δ is a blocking category for β (Δ ≠ IP) (ii) Δ is a maximal projection immediately dominating a blocking category for β (iii) Δ is the immediate projection of γ, a governor of β

Marcel den Dikken @ EGG 2016 48 Barriers — The Mechanics Q: does find govern the trace of who? → depends on one’s definition of government Chomsky’s (1986) definition of government is based on m-command — so nice governs its own specifier, and thereby blocks find from governing the trace of who so there shouldn’t be a minimality problem with the derivation below: VP ≠ a minimality barrier [CP who [C=did [IP she [VP find [AP t nice ]]]]]?

Marcel den Dikken @ EGG 2016 49 Barriers — The what Mechanics does Chomsky use clause (iii) for? barrier a domain Δ is a barrier for a relation R between α and β iff Δ dominates β and excludes α and (i) or (iii) (i) Δ is a blocking category for β (Δ ≠ IP) (ii) Δ is a maximal projection immediately dominating a blocking category for β (iii) Δ is the immediate projection of γ, a governor of β

Marcel den Dikken @ EGG 2016 50 Barriers — The what Mechanics does Chomsky use clause (iii) for? → the that-trace effect the lexical complementiser that is supposed to set up a minimality barrier, by (iii), for proper antecedent government of the initial trace of who by the intermediate trace in Spec. CP but if finite I governs its own specifier (as Chomsky assumes it does), this can’t be right! [CP who do you think [CP t′ [C=*that [IP t did it ]]]]?

Marcel den Dikken @ EGG 2016 51 Barriers — The what Mechanics does Chomsky use clause (iii) for? there are many problems with the minimalitybased approach to the that-trace effect, and with ‘rigid minimality’ in general one is the fact that no that-trace effect arises under long adjunct extraction [CP how do you think [CP t′ [C=that [IP I did it t ]]]]?

Marcel den Dikken @ EGG 2016 52 Barriers — The Mechanics Lasnik & Saito (1984) provide a technical way out from this all traces must receive a γ-mark: [+γ] for traces that satisfy the ECP, [–γ] for ones that don’t the timing of γ-marking is different for traces of argument and non-argument extraction: the latter receive their γ-mark only at LF by the time t is assigned its γ-mark the complementiser that has been deleted [CP how do you think [CP t′ [C=that [IP I did it t ]]]]?

Marcel den Dikken @ EGG 2016 53 Barriers — The Mechanics back to Who do you think (*that) did it? … though the complementiser that is also deleted from the LF-representation of the long subject extraction example, this does not ‘save’ the trace t in Spec. IP: it is assigned a γ-mark as soon as it emerges (i. e. , at S-structure), at which point that is still present and by hypothesis causes t to receive a fatal [–γ] mark [CP who do you think [CP t′ [C=*that [IP t did it ]]]]?

Marcel den Dikken @ EGG 2016 54 Barriers — The Mechanics Lasnik & Saito’s (1984) technical machinery works flawlessly (for other problems as well) but its conceptual underpinning is very weak, and at least for the ‘minimality problem’ posed by lexical complementisers it seems redundant there really is no ‘minimality effect’ imposed by the presence of a complementiser in C: the ‘that-trace effect’ cannot be derived from clause (iii) of the ECP of Chomsky (1986)

Marcel den Dikken @ EGG 2016 55 Barriers — The Mechanics clause (iii) is not helpful barrier a domain Δ is a barrier for a relation R between α and β iff Δ dominates β and excludes α and (i) or (iii) (i) Δ is a blocking category for β (Δ ≠ IP) (ii) Δ is a maximal projection immediately dominating a blocking category for β (iii) Δ is the immediate projection of γ, a governor of β

Marcel den Dikken @ EGG 2016 56 Barriers — The Mechanics there is nothing wrong, within Barriers, with the idea that a head can create a minimality problem for proper antecedent government of a phrasal trace: in Chomsky’s analysis of NPmovement, a head actually serves as the proper antecedent governor of an NP-trace but conceptually one would sooner imagine that a relationship between α and its trace would be interfered with only by an intervener of the same type as α: ‘Relativised Minimality’

Marcel den Dikken @ EGG 2016 57 Barriers — The what Mechanics does Chomsky use clause (ii) for? barrier a domain Δ is a barrier for a relation R between α and β iff Δ dominates β and excludes α and (i) or (iii) (i) Δ is a blocking category for β (Δ ≠ IP) (ii) Δ is a maximal projection immediately dominating a blocking category for β (iii) Δ is the immediate projection of γ, a governor of β

Marcel den Dikken @ EGG 2016 58 Barriers — The Mechanics the PRO-subject of a control infinitive in the complement of a verb must be ‘shielded’ from government by the verb since CP is itself L-marked (i. e. , θ-governed by V), it is not an inherent barrier for gov’t of PRO the IP in the complement of C is not L-marked but not thereby a barrier for gov’t of PRO but because IP = BC for PRO, CP inherits barrierhood for PRO from IP [IP he tried [CP C=Ø [IP PRO to win the race ]]]

Marcel den Dikken @ EGG 2016 59 Barriers — The Mechanics this account of control complements works technically, but empirical evidence for a nullheaded CP in the structure of control infinitives has never been forthcoming there alternative analyses for control infinitives on the market that do not need PRO to be ungoverned — or do not use PRO at all the analysis of control infinitives does not qualify as strong support for clause (ii) [IP he tried [CP C=Ø [IP PRO to win the race ]]]

Marcel den Dikken @ EGG 2016 60 Barriers — The Mechanics Chomsky (1986) also relies on clause (ii) in the account of the severe Subjacency effect under wh-extraction from relative clauses in the example below, CP is an inherent barrier because it is not θ-governed but the sentence is far worse than a typical ‘mild Subjacency violation’ this follows if NP inherits barrierhood from CP * [CP what did she harass [NP the [N′ man] [CP who [IP Bill was talking to t about t ]]]]]?

Marcel den Dikken @ EGG 2016 61 Barriers — The Mechanics but clause (ii) can’t help explain the deviance of CNPC effects involving so-called nouncomplement clauses here NP can’t inherit barrierhood from CP ≠ BC in fact, because CP ≠ BC, the deviance of extraction from N-complement clauses follows only by stipulation: the CP in the complement of N, though L-marked, is an inherent barrier * [CP what did she dispute [NP the [N′ claim [CP t′ [that [IP Bill was talking about t ]]]]]?

Marcel den Dikken @ EGG 2016 62 Barriers — The Mechanics in the noun-complement clause subcase of the CNPC, none of the clauses of Chomsky’s (1986) definition of a barrier delivers clause (i) fails to declare CP an inherent barrier clause (ii) is not operative because CP ≠ BC clause (iii) could possibly assign a [–γ] mark to t′ but by Lasnik & Saito’s (1984) logic this trace deletes at LF and cannot cause an ECP effect * [CP what did she dispute [NP the [N′ claim [CP t′ [that [IP Bill was talking about t ]]]]]?

Marcel den Dikken @ EGG 2016 63 Barriers — Summary Chomsky’s (1986) aim is to unify and simplify theories of trace licensing (ECP) and bounding (Subjacency) and to eliminate the stipulativeness of the latter by replacing it with an algorithmic approach to the computation of barriers the barriers algorithm is complex, consisting of three clauses, the first two of which are in turn defined in terms of the complex notion of ‘blocking category’

Marcel den Dikken @ EGG 2016 64 Barriers — Summary we have seen that the definition of ‘blocking category’ can and arguably should be simplified, making reference to θ-government only (and perhaps even further, to just the complement-of relation) clause (i), thus simplified, works reasonably well (though the exception for IP remains bad) I’s ‘defectiveness’ permeates all of Barriers (IP not a legitimate adjunction site; I′ is never a minimality barrier) — and recurs in minimalism

Marcel den Dikken @ EGG 2016 65 Barriers — Summary we have seen that the definition of ‘blocking category’ can and arguably should be simplified, making reference to θ-government only (and perhaps even further, to just the complement-of relation) clause (i), thus simplified, works reasonably well (though the exception for IP remains bad) clause (ii) is awkward and not very helpful clause (iii) does not deliver even for what it is designed to account for: that-trace effects

Marcel den Dikken @ EGG 2016 66 Minimality the type of minimality clause (iii) represents is often called ‘rigid minimality’: heads that serve as closer governors obstruct antecedent gov’t relations of all stripes, regardless of whether the antecedent is itself a head or a phrase that heads cannot engage in antecedent gov’t relations across a head nearer the trace seems clearly in evidence: the Head Mov’t Constraint What could they have been up to? *What have they could been up to?

Marcel den Dikken @ EGG 2016 67 Inflection & Verb Movement (1) (2) a. John kisses often kisses her b. John kisses not kisses often kisses her c. John does not does kiss her d. PRO kissing not kissing often e. PRO doing not doing often a. John has b. John has not has c. PRO having not having often [XP NP [X' X often kissing her doing kiss her often has kissed her having kissed her [VP Vaux [VP Vlex NP [Neg [YP Y [VP Adv Pollock: T(ense) Belletti: Agr does Agr T(ense)

Marcel den Dikken @ EGG 2016 68 Inflection & Verb Movement • → → Pollock (1989): English auxiliaries can leave VP because they have no θ-grids has, having can reach Agr without any difficulty raising Agr to T past Neg is licit, per Chomsky’s (1986) idea that an I-head receiving a verb can retroactively L-mark its complement, rendering it transparent (2) a. John has b. John has not has c. PRO having not having often [XP NP [X' X often has kissed her having kissed her [VP Vaux [VP Vlex NP [Neg [YP Y [VP Adv Pollock: T(ense) Agr

Marcel den Dikken @ EGG 2016 69 Inflection & Verb Movement • though Neg. P’s inherent barrierhood could be retroactively lifted after Agr-to-T movement, Neg. P still remains a barrier via inheritance from Agr. P (which is not L-marked), and via minimality (Neg is the closest governor of the trace in Y=Agr) → movement of Agr past Neg to T, a Head Mov’t Conbut the inheritance Pollock could have recourse straint violation, can succeed only if we abandon barrier is not eliminable here to Chomsky’s idea that clause (ii) and dodge empty clause (iii)dobynotmagic heads qualify as governors [XP NP [X' X [Neg [YP Y [VP Adv Pollock: T(ense) Agr [VP Vaux [VP Vlex NP

Marcel den Dikken @ EGG 2016 70 Inflection & Verb Movement • Chomsky (1995: Chapter 2) shifts the focus to economy of derivation — movement is not free as long as it is not forbidden (as it was in Pollock 1989) but forbidden unless forced (‘last resort’/‘least effort’) → Chomsky’s central question: how Pollock’s proposals ‘might bear on “least effort” guidelines and the status of the HMC’ (Head Movement Constraint; Travis 1984) • HMC/ECP and derivations vs representations → ‘the HMC is not a principle’: it is ‘valid only insofar as it reduces to the ECP’

Marcel den Dikken @ EGG 2016 71 Inflection & Verb Movement HMC/ECP and derivations vs representations (i) negative sentences and auxiliary movement does not violate ECP when trace is derivationally created … but does violate ECP in the LF representation T Neg → Agr ec Vt. V traces which, in the LF representation, are separated from their antecedents by an intervening head are legitimate if they are locally licensed in the course of the derivation

Marcel den Dikken @ EGG 2016 72 Inflection & Verb Movement HMC/ECP and derivations vs representations (ii) negative sentences and do-support → the dummy auxiliary do originates in Agr and raises up to T, crossing Neg (as in Pollock’s analysis) ‘ec’ is a featureless empty category, not a trace, hence not subject to ECP T violates HMC → Neg Agr=do ec Vlex … but not ECP! steps in the derivation which violate ‘shortest move’ but which yield a legitimate LF representation are permitted

Marcel den Dikken @ EGG 2016 73 Inflection & Verb Movement HMC/ECP and derivations vs representations → → → NB ‘the HMC is not a principle’: it is ‘valid only insofar as it reduces to the ECP’ traces which, in the LF representation, are separated from their antecedents by an intervening head are legitimate if they are locally licensed in the course of the derivation steps in the derivation which violate ‘shortest move’ but which yield a legitimate LF representation are permitted this is plainly an attempt to have one’s cake and eat it, too — playing fast and loose with derivations vs representations, the resulting picture is incoherent

Marcel den Dikken @ EGG 2016 74 Inflection & Verb Movement HMC/ECP and derivations vs representations → ‘the HMC is not a principle’: it is ‘valid only insofar as it reduces to the ECP’ it is fair to say that the jury is still out on the question of whether the Head Movement Constraint can be reduced to the ECP (or its successor, the Minimal Link Condition) but at a minimum, it is clear that indisputable cases of head movement crossing an overt head are ungrammatical What could they have been up to? *What have they could been up to?

Marcel den Dikken @ EGG 2016 back to… 75 Minimality the type of minimality clause (iii) represents is often called ‘rigid minimality’: heads that serve as closer governors obstruct antecedent gov’t relations of all stripes, regardless of whether the antecedent is itself a head or a phrase that heads cannot engage in antecedent gov’t relations across a head nearer the trace seems clearly in evidence: the Head Mov’t Constraint but it is not clear at all that phrasal antecedent gov’t relations are interfered with by heads

Marcel den Dikken @ EGG 2016 76 Minimality consider again the case of long adjunct mov’t [CP how do you [V′ think [CP t′ that I did it t ]]]? the intermediate trace in Spec. CP cannot be antecedent governed if V governs it and the matrix subject originates in Spec. VP (NB: not assumed in Barriers): V′ will be a MB for t′ but it is not clear at all that phrasal antecedent gov’t relations are interfered with by heads

Marcel den Dikken @ EGG 2016 77 Minimality on the other hand, phrases of the same type as the antecedent do seem to interfere with antecedent government of a phrasal trace * [CP how do you wonder [CP what I did t t ]]? but it is not clear at all that phrasal antecedent gov’t relations are interfered with by heads

Marcel den Dikken @ EGG 2016 78 Minimality if in the realm of minimality likes interfere with likes but not with non-likes, we will want to replace (iii) with a different type of minimality condition: one that codifies the ‘likeness’ that heads cannot engage in antecedent gov’t relations across a head nearer the trace seems clearly in evidence: the Head Mov’t Constraint but it is not clear at all that phrasal antecedent gov’t relations are interfered with by heads

Marcel den Dikken @ EGG 2016 79 Minimality if in the realm of minimality likes interfere with likes but not with non-likes, we will want to replace (iii) with a different type of minimality condition: one that codifies the ‘likeness’ in a configuration [. . . α. . . [. . . δ. . . [. . . β. . . ]]], minimality blocks a government relationship between α and β iff (a) δ is a head that governs β (Chomsky ’ 86) (b) δ is of the same type as α (Rizzi 1990)

Marcel den Dikken @ EGG 2016 80 Minimality if α is a head, δ is a head if α is in an A–position, δ is an A–specifier if α is in an Aʹ–position, δ is an Aʹ–specifier in a configuration [. . . α. . . [. . . δ. . . [. . . β. . . ]]], minimality blocks a government relationship between α and β iff (a) δ is a head that governs β (Chomsky ’ 86) (b) δ is of the same type as α (Rizzi 1990)

Marcel den Dikken @ EGG 2016 81 Minimality if α is a head, δ is a head if α is in an A–position, δ is an A–specifier if α is in an Aʹ–position, δ is an Aʹ–specifier for passive constructions, there is good reason to believe that the implicit external argument is syntactically represented, as a silent pronoun in the VP-internal subject position, Spec. VP (or, in the current ‘light verb’ approach, Specv. P) [IP he was [VP pro/PRO [V′ attacked t ]]]

Marcel den Dikken @ EGG 2016 82 Minimality if α is a head, δ is a head if α is in an A–position, δ is an A–specifier if α is in an Aʹ–position, δ is an Aʹ–specifier how does this derivation jibe with minimality? on its face, it doesn’t but the logic of minimality (rigid or relativised) is to make antecedent–trace relations as close as possible, avoiding potential ambiguity [IP he was [VP pro/PRO [V′ attacked t ]]]

Marcel den Dikken @ EGG 2016 83 Minimality if α is a head, δ is a head if α is in an A–position, δ is an A–specifier if α is in an Aʹ–position, δ is an Aʹ–specifier the occupant of Spec. VP is not a possible antecedent for t — chains have at most one θ-role but the logic of minimality (rigid or relativised) is to make antecedent–trace relations as close as possible, avoiding potential ambiguity [IP he was [VP pro/PRO [V′ attacked t ]]]

Marcel den Dikken @ EGG 2016 84 Minimality if α is a head, δ is a head if α is in an A–position, δ is an A–specifier if α is in an Aʹ–position, δ is an Aʹ–specifier the occupant of Spec. VP is not a possible antecedent for t — chains have at most one θ-role Guasti’s fiat: θ-positions don’t count [IP he was [VP pro/PRO [V′ attacked t ]]]

Marcel den Dikken @ EGG 2016 85 Minimality if α is a head, δ is a head if α is in an A–position, δ is an A–specifier if α is in an Aʹ–position, δ is an Aʹ–specifier BUT the problem may be much more general Guasti’s fiat: θ-positions don’t count [IP he was [VP pro/PRO [V′ attacked t ]]]

Marcel den Dikken @ EGG 2016 86 Subjects & Objects • → → subjects can agree with the finite verb this seems tied to raising to Spec. IP: these facts are the central puzzle vs. the central puzzle is these facts a lot of people are in this room vs. there is a lot of people in this room subject-Vfin agreement via Spec-Head agreement

Marcel den Dikken @ EGG 2016 87 Subjects & Objects • → objects can also agree with the verb this is not apparent in English BUT in heavily inflecting languages such as Xhosa or Quechua we see this

Marcel den Dikken @ EGG 2016 88 Subjects & Objects • → objects can also agree with the verb this is not apparent in English BUT in heavily inflecting languages such as Xhosa or Quechua we see this Xh u-mama u-ya-wu-phek-a 1 a-mother 1 a. Agr. S-Pres-3 Agr. O-cook-Asp um-ngqusho 3 -samp ‘Mother cooks samp’

Marcel den Dikken @ EGG 2016 89 Subjects & Objects • → objects can also agree with the verb this is not apparent in English BUT in heavily inflecting languages such as Xhosa or Quechua we see this Qu pro riku-wa-rqa-nki pro see-1 sg-Past-2 ‘you saw me’

Marcel den Dikken @ EGG 2016 90 Subjects & Objects Xh Qu → → → u-mama u-ya-wu-phek-a 1 a-mother 1 a-Pres-3 -cook-Asp um-ngqusho 3 -samp ‘Mother cooks samp’ pro riku-wa-rqa-nki pro see-1 sg-Past-2 ‘you saw me’ Xhosa is almost consistently prefixing Quechua is consistently suffixing in both, object agreement is closer to the stem than subject agreement; T in between

Marcel den Dikken @ EGG 2016 Subjects & Objects Xh Qu u-mama u-ya-wu-phek-a 1 a-mother 1 a-Pres-3 -cook-Asp um-ngqusho 3 -samp ‘Mother cooks samp’ pro riku-wa-rqa-nki pro see-1 sg-Past-2 ‘you saw me’ Xhosa: Agr. S-T-Agr. O-Vstem Quechua: Vstem-Agr. O-T-Agr. S 91

Marcel den Dikken @ EGG 2016 Subjects & Objects Xh Qu u-mama u-ya-wu-phek-a 1 a-mother 1 a-Pres-3 -cook-Asp um-ngqusho 3 -samp ‘Mother cooks samp’ pro riku-wa-rqa-nki pro see-1 sg-Past-2 ‘you saw me’ Xhosa: Agr. S-T-Agr. O-Vstem Quechua: Vstem-Agr. O-T-Agr. S → this is NOT an accident 92

Marcel den Dikken @ EGG 2016 Subjects & Objects • • assimilating subject agreement and object agreement splitting IP into projections for – subject agreement (Agr. S) – object agreement (Agr. O) – tense (T) (Chomsky 1991 = Chomsky 1995: Chapter 2, extending Pollock 1989) 93

Marcel den Dikken @ EGG 2016 Subjects & Objects • → • subjects agree with the finite verb by raising to Spec. Agr. SP subject-verb agreement via Spec-Head agreement in Agr. SP objects agree with the finite verb by raising to Spec. Agr. OP object-verb agreement via Spec-Head agreement in Agr. OP TP in between Agr. SP and Agr. OP 94

Marcel den Dikken @ EGG 2016 95 Agr. SP SUj Agr. S' Agr. S TP T' T Agr. OP OBi Agr. O' Agr. O VP tj V V' ti the morphological make-up mirrors the syntactic derivation (with variation residing in whether the morphology is prefixal or suffixal)

Marcel den Dikken @ EGG 2016 96 Agr. SP SUj Agr. S' Agr. S TP T' T Xh Agr. OP OBi Agr. O' Agr. O VP tj V u-ya-wu-phek-a 1 a. Agr. S-Pres-3 Agr. O-cook-Asp V' ti

Marcel den Dikken @ EGG 2016 97 Agr. SP SUj Agr. S' Agr. S TP T' T Qu riku-wa-rqa-nki see-1 sg-Past-2 Agr. OP OBi Agr. O' Agr. O VP tj V V' ti

Marcel den Dikken @ EGG 2016 98 Agr. SP SUj Agr. S' Agr. S TP T' T Agr. OP OBi Agr. O' Agr. O VP tj V V' ti with both subjects and objects in spec positions, we can now assimilate Case assignment to subjects and objects via Spec-Head agreement

Marcel den Dikken @ EGG 2016 99 Agr. SP SUj Agr. S' Agr. S TP T' T Agr. OP OBi Agr. O' Agr. O VP tj V V' Non-trivial Question: How does this derivation obey Shortest Move (economy of derivation)? ti

Marcel den Dikken @ EGG 2016 100 Step 1: OB-movt Agr. SP Agr. S' Agr. S TP T' T Agr. OP Guasti’s fiat: OBi Agr. O' θ-positions Agr. O VP don’t count SU V violates Shortest Move? V' ti perhaps we can simply ignore the subject’s base position?

Marcel den Dikken @ EGG 2016 101 Step 2: SU-movt Agr. SP SUi Agr. S' Agr. S TP T' T violates Shortest Move? Agr. OP OBi Agr. O' Agr. O VP SU V V' the closest potential landing-site for SU-movt is Spec. TP ti

Marcel den Dikken @ EGG 2016 102 Step 2: SU-movt Agr. SP SUi Agr. S' Agr. S TP T' T violates Shortest Move? Agr. OP OBi Agr. O' Agr. O VP SU V V' now we can’t ignore the object’s landing site: this is a θ′-position, so it does count ti

Marcel den Dikken @ EGG 2016 103 the desired derivation Agr. SP SUj Agr. S' Agr. S ‘crossing paths’ TP tj T' T Agr. OP OBi Agr. O' Agr. O VP tj V V' Spec. VP and Spec. Agr. OP must be ‘equidistant’ Spec. Agr. OP and Spec. TP must be ‘equidistant’ ti

Marcel den Dikken @ EGG 2016 * the non-desired derivation Agr. SP OBi 104 Agr. S' Agr. S ‘nested paths’ TP ti T' T Agr. OP SUj Agr. O' Agr. O VP tj V V' ti

Marcel den Dikken @ EGG 2016 * 105 for this derivation to work, these must all be equidistant Agr. SP OBi Agr. S' Agr. S TP ti T' T Agr. OP SUj Agr. O' Agr. O VP tj V V' Key Question: How to rule in the ‘crossing paths’ derivation, while ruling out the ‘nested paths’ derivation? ti

Marcel den Dikken @ EGG 2016 106 Shortest Move & Equidistance episode I overt movement of both NPs in a transitive clause: the crossing paths derivation

Marcel den Dikken @ EGG 2016 107 Agr. SP SUj Agr. S' Agr. S TP tj T' T Agr. OP OBi Agr. O' Agr. O VP tj V V' ti

Marcel den Dikken @ EGG 2016 108 Agr. SP Agr. S' Agr. S TP T' T must be equidistant Agr. OP OBi Agr. O' Agr. O VP SU V V' ti

Marcel den Dikken @ EGG 2016 minimal Agr. SP domain of X DMIN (X) = SU Agr. S' i all nodes Agr. S dominated TP by XP minus X, any nodes dominatingti X, T and any nodes dominated by members of DMIN(X) must be equidistant 109 {Spec. XP, compl-of-X} T' Agr. OP OBi Agr. O' Agr. O VP SU V V' ti to get these two specifiers equidistant, we must find some (i. e. , any) minimal domain that contains both positions to get two specifiers in one DMIN, we must perform head movement: no stationary head ever has a DMIN w/ two specs

Marcel den Dikken @ EGG 2016 minimal Agr. SP domain of chain DMIN (X, Y) = SU Agr. S' i all nodes Agr. S dominated by. TP XP minus X and Y, any nodes dominatingti X or T Y, and any nodes dominated by members of DMIN(X, Y) 110 {Spec. XP, Spec. YP, compl-of-Y} T' [X=head, Y=foot of CH] Agr. OP Agr. O' Agr. O VP SU V STEP 1: the verb raises up to Agr. O STEP 2: Merge Spec. Agr. OP and Spec. VP (= SU) are equidistant V' OB

Marcel den Dikken @ EGG 2016 111 Agr. SP SUi Agr. S' Agr. S TP ti T' T Agr. OP Agr. O' Agr. O VP SU V V' OB STEP 2: Merge Spec. Agr. OP STEP 3: OB raises to Spec. Agr. OP, obeying Shortest Move

Marcel den Dikken @ EGG 2016 112 Agr. SP SUi Agr. S' Agr. S TP ti T' T ti the result of STEPs 1 -3 Agr. OP OBi Agr. O' Agr. O VP SU V Agr. O V' t. V

Marcel den Dikken @ EGG 2016 113 Agr. SP SUi Agr. S' Agr. S TP T' T Agr. OP OBi Agr. O' Agr. O VP SU t. V V Agr. O next, SU wants to raise the closest potential landing-site is Spec. TP V' ti

Marcel den Dikken @ EGG 2016 102 Agr. SP SUi Agr. S' Agr. S TP T' T Agr. OP OBi Agr. O' Agr. O VP SU t. V V Agr. O STEP 4: Agr. O raises up to T STEP 5: Merge Spec. TP and Spec. Agr. OP are equidistant V' ti

Marcel den Dikken @ EGG 2016 115 Agr. SP SUi Agr. S' Agr. S TP T' T Agr. OP OBi Agr. O' Agr. O VP SU t. V V Agr. O V' STEP 5: Merge Spec. TP STEP 6: SU raises to Spec. TP, obeying Shortest Move ti

Marcel den Dikken @ EGG 2016 116 Agr. SP SUi Agr. S' Agr. S SUj TP T' T Agr. OP OBi Agr. O' Agr. O T t. Agr. O VP tj V Agr. O t. V the result of STEPs 4 -6 V' ti

Marcel den Dikken @ EGG 2016 117 Agr. SP SUi Agr. S' Agr. S TP ti T' T Agr. OP OBj Agr. O' t. Agr. O VP ti t. V V' tj finally SU raises from Spec. TP to Spec. Agr. OP (local mov’t); this is not contingent on domain-extending head movement; T eventually raises to Agr. S (not necessarily in overt syntax)

Marcel den Dikken @ EGG 2016 118 Shortest Move & Equidistance episode II overt movement of both NPs in a transitive clause: the nested paths derivation

Marcel den Dikken @ EGG 2016 119 recall first of all the crossing paths story Agr. SP SUj Agr. S' Agr. S TP tj T' T Agr. OP OBi Agr. O' Agr. O VP tj V V' ti

Marcel den Dikken @ EGG 2016 * 120 now consider the nested paths story Agr. SP OBi Agr. S' Agr. S TP ti T' T Agr. OP SUj Agr. O' Agr. O VP tj V V' ti

Marcel den Dikken @ EGG 2016 * 121 Agr. SP OBi cannot all be equidistant Agr. S' Agr. S TP ti T' T Agr. OP SUj Agr. O' Agr. O VP tj V V' ti

Marcel den Dikken @ EGG 2016 * 122 Agr. SP OBi in DMIN of CH (V, t) Agr. S' Agr. S TP ti T' T Agr. OP SUj Agr. O' Agr. O VP tj V V' ti

Marcel den Dikken @ EGG 2016 * 123 Agr. SP OBi in DMIN of CH (Agr. O, t) Agr. S' Agr. S TP ti T' T Agr. OP SUj Agr. O' Agr. O VP tj V V' ti

Marcel den Dikken @ EGG 2016 * 124 Agr. SP OBi Agr. S' Agr. S TP ti T' T Agr. OP SUj Agr. O' Agr. O VP tj V V' ti

Marcel den Dikken @ EGG 2016 * 125 Agr. SP OBi head movement chains have at most two members → CH (V, t, t) is ill-formed Agr. S' Agr. S TP ti T' T Agr. OP SUj Agr. O' Agr. O VP tj V V' ti

Marcel den Dikken @ EGG 2016 * 126 Agr. SP OBi head movement also can’t skip intermediate heads: this would violate locality Agr. S' Agr. S TP ti T' T Agr. OP SUj Agr. O' Agr. O VP tj V V' ti

Marcel den Dikken @ EGG 2016 127 Shortest Move & Equidistance episode III overt movement of both NPs in a transitive clause: the reordered crossing paths derivation

Marcel den Dikken @ EGG 2016 128 recall once again the episode I derivation Agr. SP SUj Agr. S' Agr. S TP tj T' T Agr. OP OBi Agr. O' Agr. O VP tj V V' OB moves to Spec. Agr. OP before SU moves up the tree ti

Marcel den Dikken @ EGG 2016 * 129 now consider the episode III derivation Agr. SP SUj Agr. S' Agr. S TP tj T' T Agr. OP OBi Agr. O' Agr. O VP tj V V' OB moves to Spec. Agr. OP after SU moves up the tree ti

Marcel den Dikken @ EGG 2016 130 Agr. SP Agr. S' TP Agr. S T' tj T Agr. O' Agr. O VP SU tj j V V' what do we do next? … depends on whether OB moves ti i OB

Marcel den Dikken @ EGG 2016 131 Agr. SP Agr. S' TP Agr. S T' tj T #1: OB moves… Agr. OP Agr. O' Agr. O VP SU tj j V … so steps 1 -3 ensue V' ttii i OB

Marcel den Dikken @ EGG 2016 132 Agr. SP Agr. S' TP Agr. S tj Agr. O' Agr. O VP SU tj j V V' ti i OB #2: OB doesn’t move… … so no need to merge Spec. Agr. OP

Marcel den Dikken @ EGG 2016 133 Agr. SP Agr. S' TP Agr. S tj Agr. OP Agr. O' Agr. O VP SU tj j V V' ti i OB #2: OB doesn’t move… … we project vacuously to Agr. OP

Marcel den Dikken @ EGG 2016 134 Agr. SP Agr. S' TP Agr. S T' tj T Agr. OP Agr. O' Agr. O VP SU tj j V V' ti i OB #2: OB doesn’t move… … we then proceed to merging T

Marcel den Dikken @ EGG 2016 135 Agr. S' TP Agr. S T' tj T Agr. OP Agr. O' Agr. O VP SU tj j V … and so on up to Agr. S' V' ti i OB

Marcel den Dikken @ EGG 2016 136 Agr. SP Agr. S' TP Agr. S T' tj T Agr. OP Agr. O' Agr. O VP SU ttjj j V V' … at which point SU can move … with no need for head mov’t! ti i OB

Marcel den Dikken @ EGG 2016 137 Agr. SP SUj Agr. S' TP Agr. S T' tj T Agr. OP Agr. O' Agr. O VP ttjj V V' ti i OB … after which OB can move … with just V-to-Agr. O being needed

Marcel den Dikken @ EGG 2016 * 138 Agr. SP SUj the episode III derivation in summarised format Agr. S' Agr. S TP T' T Agr. OP OBi Agr. O' Agr. O VP tj V this needs to be blocked → if this were legitimate, Holmberg’s Generalisation would fail to be derived! V' ti

Marcel den Dikken @ EGG 2016 Shortest Move & Equidistance The Extension Condition any application of Move/Merge must extend the tree at the root (i. e. , ‘must build the tree taller’) 139

Marcel den Dikken @ EGG 2016 140 Agr. SP SUj ‘OB after SU’ derivation revisited Agr. S' TP Agr. S T' tj T Agr. OP Agr. O' Agr. O VP ttjj V V' ti i OB OB-movement here violates the Extension Condition … because merging Spec. Agr. OP does not extend the tree

Marcel den Dikken @ EGG 2016 141 Agr. SP SUi recall the episode I derivation Agr. S' Agr. S TP … the Extension Condition is. T'satisfied t i T Agr. OP Agr. O' Agr. O VP SU V STEP 1: the verb raises up to Agr. O STEP 2: Merge Spec. Agr. OP STEP 3: OB raises to Spec. Agr. OP V' OB

Marcel den Dikken @ EGG 2016 147 Shortest Move & Equidistance episode IV overt movement of SU, covert movement of OB … the case of English

Marcel den Dikken @ EGG 2016 148 Agr. SP SUj the covert portion of the English-type derivation… Agr. S' TP Agr. S T' tj T Agr. OP Agr. O' Agr. O VP ttjj V LF-movement (‘covert’) of OB to Spec. Agr. OP … violates the Extension Condition … but the derivation must converge … so we are led to conclude that … V' ti i OB

Marcel den Dikken @ EGG 2016 Shortest Move & Equidistance The Extension Condition any application of Move/Merge must extend the tree at the root (i. e. , ‘must build the tree taller’) … does not hold at LF 149

Marcel den Dikken @ EGG 2016 150 Shortest Move & Equidistance The Extension Condition any application of Move/Merge must extend the tree at the root (i. e. , ‘must build the tree taller’) … which makes it an S-structure condition!

Marcel den Dikken @ EGG 2016 151 What determines when a phrase moves? feature strength – take 1 (Chapter 3) overt movement is triggered by a strong feature on a functional head; weak features do not trigger overt mov’t … because (i) weak FFs are PF-invisible, and (ii) it is better to move covertly than overtly

Marcel den Dikken @ EGG 2016 152 What determines when a phrase moves? feature strength – take 2 (Chapter 4) overt movement is triggered by a strong feature on a functional head; weak features do not trigger overt mov’t … because (i) strong FFs cannot be tolerated; (ii) it is better to move covertly than overtly

Marcel den Dikken @ EGG 2016 153 What does it mean for a feature to be ‘intolerable’? a feature that the derivation cannot tolerate must be checked (i. e. , eliminated) before the structure is built taller … i. e. , before the F-head that bears it is included in a larger structure that no longer bears F’s label

Marcel den Dikken @ EGG 2016 Agr. SP back to the tree … Agr. S' TP Agr. S tj T 154 T' Agr. O VP <FF> SU tj j V V' ti i OB <FF> what do we do next? » if FF on Agr. O is strong…

Marcel den Dikken @ EGG 2016 155 Agr. SP back to the tree … Agr. S' TP Agr. S tj T T' Agr. O VP <FF> SU tj j V V' ti i OB <FF> … then FF must be checked at once … and OB<FF> raises to Spec. Agr. OP

Marcel den Dikken @ EGG 2016 Agr. SP back to the tree … Agr. S' TP Agr. S tj T 156 T' Agr. OP Agr. O' Agr. O VP <FF> SU tj j V V' ttii i OB <FF>

Marcel den Dikken @ EGG 2016 Agr. SP back to the tree … Agr. S' TP Agr. S tj T 157 T' Agr. O VP <FF> SU tj j V V' ti i OB <FF> what do we do next? » if FF on Agr. O is weak …

Marcel den Dikken @ EGG 2016 158 Agr. SP back to the tree … Agr. S' TP Agr. S tj T T' Agr. O VP <FF> SU tj j V V' ti i OB <FF> … then no need to check FF now … so we proceed to merging T, etc.

Marcel den Dikken @ EGG 2016 recall… 159 the Extension Condition is built into the definition of strong features → the Extension Condition can go, and with it the apparent S-structure constraint! a feature that the derivation cannot tolerate must be checked (i. e. , eliminated) before the structure is built taller … i. e. , before the F-head that bears it is included in a larger structure that no longer bears F’s label

Marcel den Dikken @ EGG 2016 160 Results • subject and object check agreement in the same way: SHAGR • subject and object check Case in the same way: SHAGR ☻ government is redundant in the domain of Case Theory

Marcel den Dikken @ EGG 2016 161 Results • subject and object do not necessarily establish a SHAGR relation with their AGR-heads at the same point in the derivation (a) Scandinavian: SU-to-Agr. SP overt; OB-to-Agr. OP may be overt, too (b) English: SU-to-Agr. SP overt; OB-to-Agr. OP covert (c) (some) VSO lgs: SU-to-Agr. SP covert; OB-to-Agr. OP covert

Marcel den Dikken @ EGG 2016 162 Results • • movement of SU and OB is triggered by the need to check features the Principle of Full Interpretation demands that uninterpretable features be checked (eliminated) by LF Full Interpretation the interpretive levels (LF, PF) must not contain superfluous (uninterpretable) symbols ☺ economy of representation derives the need for feature-checking mov’t

Marcel den Dikken @ EGG 2016 163 Results • • • ☺ overtness of SU and OB movement is a function of feature strength the ‘intolerability’ of strong features forces movement to take place overtly in the absence of a compelling reason to move overtly, stay put Procrastinate: move as late as possible economy of derivation in conjunction with ‘intolerability’ of strong features derives the distribution of overt mov’t

Marcel den Dikken @ EGG 2016 164 Results ☺ • ☻ economy and the ‘intolerability’ of strong features derive all there is to say about movement and its timing with Case analysed as a [–int] FF that needs to be checked, economy also derives the Case Filter and the distribution of (c)overt Case-driven mov’t Case Theory need not and cannot (recall English and VSO-lgs!) make reference to S-structure

Marcel den Dikken @ EGG 2016 165 Results • • the ‘intolerability’ of strong features entails that movement can proceed before the structure is fully built Merge (structure building) and Move can take turns: Move does not ‘wait’ until the entire D-structure is erected, but applies as soon as it must

Marcel den Dikken @ EGG 2016 Agr. SP back to the tree … Agr. S' TP Agr. S tj T 166 T' Agr. O VP <FF> SU tj j V V' ti i OB <FF> what do we do next? » if FF on Agr. O is strong…

Marcel den Dikken @ EGG 2016 167 Agr. SP back to the tree … Agr. S' TP Agr. S tj T T' Agr. O VP <FF> SU tj j V V' ti i OB <FF> … then FF must be checked at once … and OB<FF> raises to Spec. Agr. OP

Marcel den Dikken @ EGG 2016 Agr. SP back to the tree … Agr. S' TP Agr. S tj T 168 T' Agr. OP Agr. O' Agr. O VP <FF> SU tj j V V' ttii i OB <FF> before T is merged!

Marcel den Dikken @ EGG 2016 169 Results • • ☻ the ‘intolerability’ of strong features entails that movement can proceed before the structure is fully built Merge (structure building) and Move can take turns: Move does not ‘wait’ till the entire D-structure is erected, but applies as soon as it must reference to full-fledged D-structure representations as input to Move is impossible

Marcel den Dikken @ EGG 2016 170 ‘Split Infl’ Q 1: Could we simplify the ‘split-IP’ structure? Could we manage without the Agr. Ps? Q 2: Could we force IP structure to be simplified? Could the Agr. P structure be argued to be bad? NB 1: ‘agreement’ is a relationship, not in any obvious sense a node in the tree NB 2: ‘Agr’ qua node is totally devoid of interpretation

Marcel den Dikken @ EGG 2016 ‘Split Infl’ Agr. SP SUj 171 Agr. S' Agr. S sample derivation TP tj (recapitulation) T' T Agr. OP OBi Agr. O' Agr. O VP tj V V' a bit of an ‘embarrassment of riches’ upstairs… ti

Marcel den Dikken @ EGG 2016 ‘Split Infl’ Agr. SP SUj 172 Agr. S' Agr. S sample derivation TP tj (recapitulation) T' T Agr. OP OBi Agr. O' Agr. O VP tj V V' … we don’t seem to need both Agr. SP and TP ti

Marcel den Dikken @ EGG 2016 173 ‘Little v’ the alternative TP SUj (Chomsky 1995: Ch. 4) T' T v. P OBi v. P tj v' v VP V the v is a agentive/causative ‘light verb’ ti

Marcel den Dikken @ EGG 2016 174 ‘Little v’ Chomsky (1995: Ch. 4): the external argument is not introduced by V but by a ‘light verb’ v projecting outside VP domain for V and int. arg. v. P ‘light verb’ ext. arg. v′ v VP V

Marcel den Dikken @ EGG 2016 175 ‘Little v’ v introduces the external argument v categorises the root ‘V’ as verbal (Distributed Morphology) v licenses a structural accusative Case feature <ACC> Burzio’s Generalisation is ‘embodied by’ v v. P ext. arg. v′ v VP V

Marcel den Dikken @ EGG 2016 176 ‘Little v’ the alternative TP SUj (Chomsky 1995: Ch. 4) T' T v. P OBi v. P tj v' v … contra Kayne’s (1994) antisymmetry VP V v needs multiple specifiers to play its part ti

Marcel den Dikken @ EGG 2016 Rethinking X-bar Theory Chomsky (1995: Chapter 3) does not quibble with X-bar Theory On the contrary, section 2 of chapter 3 is entitled: ‘Fundamental Relations: X-Bar Theory’ 177

Marcel den Dikken @ EGG 2016 178 Rethinking X-bar Theory recognises (at least) one distinct projection level in addition to the head (X 0) and maximal projection (XP) — the X′ level note, however, that X′ is an intermediate projection level, never addressed directly by any operation or principle of the grammar → → → but note that in Barriers, X′ (esp. C′) is argued to be a Minimality barrier X′ cannot be selected X′ cannot be moved X′ cannot be deleted/elided/replaced

Marcel den Dikken @ EGG 2016 179 Rethinking X-bar Theory it may not be obvious that X′ cannot be deleted/elided/ replaced, in view of such examples as (1) and (2) (1) a. b. you’ve got your troubles, I’ve got mine ec [NP mine [N′ ec]] (2) a. b. you like this book, I like that one [NP that [N′ one]] → → → X′ cannot be selected X′ cannot be moved X′ cannot be deleted/elided/replaced

Marcel den Dikken @ EGG 2016 180 Rethinking X-bar Theory but there is good reason, independently, to believe that noun phrases have a more elaborate structure than the one we have worked with so far a structure like (3) yields (4 a, b) for (1 a), (2 a) (1) a. b. you’ve got your troubles, I’ve got mine ec [NP mine [N′ ec]] (2) a. b. you like this book, I like that one [NP that [N′ one]] (3) (4) [DP Spec [D′ D [NP … N … ]]] a. b. [DP mine [D′ D [NP ec]]] [DP [D′ that [NP one]]]

Marcel den Dikken @ EGG 2016 181 Rethinking X-bar Theory Q 1: Could we allow X-bar structure to be simplified? Could we manage without the X'/XP distinction? Q 2: Could we force X-bar structure to be simplified? Could the X'/XP distinction be shown to be bad?

Marcel den Dikken @ EGG 2016 Rethinking X-bar Theory Agr. SP Agr. S' TP Agr. S trivial with standard X-bar T' tj theory: the NP-labelled *NP IP structure is ill-formed →T I-projection is incomplete I' IP not immediately obvious I VP with simplified X-bar NP tj i theory: I-projection is V certainly not incomplete 182 V' VP ti NP NP raises to Spec. IP Q: how do we ensure that the result is IP, not NP? I’s strong feature that triggers NP’s mov’t must be checked before I is included in a larger structure with a different label

Marcel den Dikken @ EGG 2016 183 X-bar. Agr. SP Structure, Projection, and Antisymmetry Agr. S' reason to prefer (5) to (6)? is there a compelling NO → (6)Agr. S is simpler, TP hence preferred cet. paribus T' (6) tj (5) IP IP T I' IP NP NP I VP N' tj NP V' NP VP N ti V NP N N' N N N

Marcel den Dikken @ EGG 2016 184 X-bar. Agr. SP Structure, Projection, and Antisymmetry Agr. S' reason to prefer (6) to (5)? is there a compelling Agr. S YES → (5) does not TP translate into linear order T' (6) tj (5) IP IP T I' IP NP NP I VP N' tj NP V' NP VP N ti V NP N N' N N → the antisymmetry of N syntax (Kayne 1994)

Marcel den Dikken @ EGG 2016 185 X-bar Structure, Projection, and Antisymmetry Q: A: → what is the relation between syntactic hierarchy and linear order? syntactic hierarchy ideally translates seamlessly into linear order — w/ c-command as translator how do we translate a syntactic hierarchical structure into a linear string of words (terminals)? not directly (because the terminals themselves have no syntactic relationships with each other) first compile the complete set of ordered pairs of non-terminals such that, for each pair, the first member asymmetrically c-commands the second

Marcel den Dikken @ EGG 2016 186 X-bar. Agr. SP Structure, Projection, and Antisymmetry set of ordered pairs of non-terminals Agr. S' K c-commands nothing {<J, M>, <J, N>, <J, P>, <L, I>, <M, P>} J asymmetrically c-commands TP Agr. S M, N, P T' L asymmetrically c-commands K I T M asymmetrically c-commands L J P N M N asymmetrically c-commands I nothing m P P c-commands nothing i I c-commands nothing p → first compile the complete set of ordered pairs of non-terminals such that, for each pair, the first member asymmetrically c-commands the second

Marcel den Dikken @ EGG 2016 187 X-bar. Agr. SP Structure, Projection, and Antisymmetry set of ordered pairs of non-terminals Agr. S' {<J, M>, <J, N>, <J, P>, <L, I>, <M, P>} TP set of ordered pairs of terminals T'{<i, m>, <i, p>, <m, i>, <m, p>} K T L J N M I m P i p → next compile the complete set of ordered pairs of terminals corresponding to the set of ordered pairs of non-terminals previously compiled

Marcel den Dikken @ EGG 2016 188 X-bar. Agr. SP Structure, Projection, and Antisymmetry Agr. S' TP How can we fix this? T' IP T I′ NP VP I N will V John laugh → → the tree we just condemned corresponds to the familiar X-bar theoretic structure of the clause this tree must now be rejected

Marcel den Dikken @ EGG 2016 189 X-bar. Agr. SP Structure, Projection, and Antisymmetry set of ordered pairs of non-terminals Agr. S' upper L c-commands nothing {<J, M>, <J, N>, <J, P>, <M, P>} lower L c-commands nothing TP Agr. S c-command is L category c-commands nothing T' defined for J asymmetrically c-commands categories L T M, N, P M asymmetrically c-commands L J P N M N asymmetrically c-commands I nothing m P P c-commands nothing i I c-commands nothing p → first compile the complete set of ordered pairs of non-terminals such that, for each pair, the first member asymmetrically c-commands the second

Marcel den Dikken @ EGG 2016 190 X-bar. Agr. SP Structure, Projection, and Antisymmetry set of ordered pairs of non-terminals Agr. S' {<J, M>, <J, N>, <J, P>, <M, P>} TP linear order of terminals set of ordered pairs of terminals T'{<i, m>, <i, p>, <m, p>} i<m<p L T L J N M I m P i p → next compile the complete set of ordered pairs of terminals corresponding to the set of ordered pairs of non-terminals previously compiled

Marcel den Dikken @ EGG 2016 191 X-bar. Agr. SP Structure, Projection, and Antisymmetry Agr. S' TP T' IP T IP NP VP I N will V John laugh → → the tree we just accepted corresponds to the simplified X-bar theoretic structure, lacking X′ this tree is forced upon us to ensure linearisation

Marcel den Dikken @ EGG 2016 192 X-bar. Agr. SP Structure, Projection, and Antisymmetry Agr. S' reason to prefer (6) to (5)? is there a compelling Agr. S YES → (5) does not TP translate into linear order T' (6) tj (5) IP IP T I' IP NP NP I VP N' tj NP V' NP VP N ti V NP N N' N N → the antisymmetry of N syntax (Kayne 1994)

Marcel den Dikken @ EGG 2016 193 Rethinking X-bar Theory Chomsky (1995: Chapter 4): this is definitely useful… but it does not go far enough — we need to be bolder Q 1: Could we allow X-bar structure to be simplified? Could we manage without the X'/XP distinction? Kayne asked… Q 2: Could we force X-bar structure to be simplified? Could the X'/XP distinction be shown to be bad?

Marcel den Dikken @ EGG 2016 194 Rethinking X-bar Theory Chomsky (1995: Chapter 4): this is definitely useful… but it does not go far enough — we need to be bolder Q 1: Could we allow X-bar theory to be abandoned? Could we manage without bar-level distinctions? Chomsky asks… Q 2: Could we force X-bar theory to be abandoned? Could bar-level distinctions be proven wrong? → Chomsky’s ‘Bare Phrase Structure’ attempts to show that X-bar theory is unnecessary and has to be abandoned, to allow multiple specs

Marcel den Dikken @ EGG 2016 195 Bare Phrase Structure the alternative T SUj à la ‘Bare Phrase Structure’ T T v OBi v tj but in Chomsky’s most recent work labels play a key role, and so does the head/phrase distinction v v V V ti perhaps even the labels are superfluous (Collins)