ON VARIATION IN THE POSITIONING OF THE PROGRESSIVE MARKER AM IN NON-STANDARD GERMAN: ON VARIATION IN THE POSITIONING OF THE PROGRESSIVE MARKER AM IN NON-STANDARD GERMAN: KÖLSCH (COLOGNESE) AND PENNSYLVANIA DUTCH Robert Cirillo * Abstract: Progressive constructions involving the progressive marker am are considered to be highly colloquial in Standard German but are standard in regional varieties of German such as Kölsch (Colognese) and Pennsylvania Dutch. It was argued in Bhatt and Schmidt (1993) that in Kölsch the progressive particle am is the head of a head-final Aspect Phrase. The progressive particle in Pennsylvania Dutch has not been discussed in generative literature. Based on Bhatt and Schmidt’s arguments, it will be argued here that in Pennsylvania Dutch the progressive marker am is the head of a head-initial Aspect Phrase, despite a possible conflict with the Final-Over-Final Constraint (FOFC) proposed in Biberauer et al. (2007, 2010). Keywords: progressive, German, Pennsylvania Dutch, Kölsch 1. Introduction Languages like English, Italian and Spanish have a progressive construction that involves an auxiliary that is a form of the verb be. In English the auxiliary is combined with a present participle while in Italian and Spanish it appears with a gerundive: (1) a. The girl is working. b. La ragazza sta lavorando. (Italian) the girl is working c. La muchacha está trabajando. (Spanish) the girl is working Standard German does not possess a comparable construction. However, in certain regional varieties of Standard German there is a progressive construction, considered to be highly colloquial, that consists of the progressive marker am followed by an infinitive: (2) Das Mädchen ist das Auto in der Garage am reparieren. the girl is the car in the garage PROG repair-INF ‘The girl is repairing the car in the garage.’ Notice that the progressive marker immediately precedes the infinitive and follows all other elements in VP, including objects and prepositional phrases. If the construction contains a particle verb, the progressive marker immediately precedes the particle. This is seen in the following example, in which the progressive marker am appears before the verbal particle auf ‘up’: * University of Amsterdam, robert.john.cirillo@planet.nl. https://mail.google.com/mail/u/0/h/gcme3dhr5ueu/?&cs=wh&v=b&to=robert.john.cirillo@planet.nl 6 R o b e r t C i r i l l o (3) Peter ist die Wäsche im Garten am aufhängen. Peter is the laundry in the garden PROG up hang-INF ‘Peter is hanging up the laundry in the garden.’ Because the am-progressive is mainly associated with the Rhineland, especially the city of Cologne, it is often called the rheinische Verlaufsform ‘progressive of the Rhineland’. I should point out, however, that the Rhineland does not have a monopoly on the am-progressive. It can be heard as far south as Swabia and Bavaria and as far east as Saxony. In Kölsch or Colognese, the language spoken in and around the city of Cologne, the am-progressive can be considered standard (non-colloquial). One of the ways in which it differs from the am-progressive in colloquial Standard German is that it allows am to come between a verb and a particle. Compare the following sentences from Kölsch to the example in (3) above 1 : (4) a. D’r Pitter es de Wäsch em Jade am ophange the Pitter is the laundry in the garden PROG up hang-INF ‘Peter is hanging up the laundry in the garden.’ b. D’r Pitter es de Wäsch em Jade op am hange the Pitter is the laundry in the garden up PROG hang-INF ‘Peter is hanging up the laundry in the garden.’ The word order in (4b), in which the progressive marker am comes between the verb and the verbal particle, is not possible in colloquial Standard German. Historically, the particle am is a contraction of the preposition an ‘on’ and dem, which is the dative singular neuter form of the definite article. However, in the examples in (2), (3) and (4) it would not be possible to replace am with the non-contracted form an dem. This indicates that am has been reanalyzed or grammaticalized as a progressive marker. Interestingly, in Dutch there is an almost identical construction, which is considered to be standard (non-colloquial), involving the preposition aan ‘on’ and the singular neuter form of the definite article het. Furthermore, like Kölsch and unlike colloquial Standard German, Dutch allows the progressive marker to come between a verb and a verbal particle in some circumstances. The following Dutch sentences are the equivalent of the examples in (4): (5) a. Piet is de was in de tuin aan het ophangen. Piet is the laundry in the garden PROG up hang-INF ‘Peter is hanging up the laundry in the garden.’ b. Piet is de was in de tuin op aan het hangen. Piet is the laundry in the garden up PROG hang-INF ‘Peter is hanging up the laundry in the garden.’ 1 Unless otherwise indicated, all example sentences from Kölsch are the author’s own. All examples have been cleared with at least one native speaker. On variation in the positioning of the progressive marker am in non-standard German 7 The question that immediately comes to mind is what syntactic category am belongs to. It is argued in Bhatt and Schmidt (1993) that in Kölsch am is the head of a head-final Aspect Phrase. The main purpose of this article is to consider whether Bhatt and Schmidt’s analysis is also applicable to the am-progressive found in Pennsylvania Dutch, in which am occurs in a different position. This article is organized as follows: Section 2 presents a summary of the analysis of am-progressives in Bhatt and Schmidt (1993). In Section 3, the am-progressive in Pennsylvania Dutch is introduced and an analysis is presented. Section 4 contains the main findings. 2. The am-progressive in Kölsch as analyzed in Bhatt and Schmidt (1993) Bhatt and Schmidt (1993) liken the am-progressive in Kölsch to the English progressive with the participial suffix –ing. The trees in (6) correspond to the way in which the Kölsch sentences in (4) and the English example in (1a) would be represented under Bhatt and Schmidt’s analysis. Note that the verbal domain is head-final in Kölsch (or German) and head-initial in English. In both languages there is an Auxiliary Phrase dominating AspP. The AuxP is headed by be in English and its equivalent sein ‘be’ in German. In Kölsch the progressive marker am is the head of AspP, as in (6a), while in English AspP contains a null-head and the suffix -ing is simply inflection triggered by an agreement operation within AspP 2 , as in (6b). (6) a. AuxP b. AuxP 3 3 Aux ́ Aux´ 3 3 AspP Aux Aux AspP 3 ist is 3 Asp ́ (is) Asp´ 3 3 VP Asp Asp VP 3 am  3 V ́ V´ 3 2 V V ophange work-ing (up hang) 2 In English one could of course dispense with the AspP with a null head and treat the auxiliary be as the head of AspP or ProgP. Bhatt and Schmidt (1993) presumably posit the extra layer in English for the sake of symmetry between English and German. Since this topic is only indirectly related to this article, which is about German, it will not be pursued any further here. 8 R o b e r t C i r i l l o In (6a), the entire verb ophange can move head-to-head and right adjoin to the progressive marker am, producing (4a). However, the verb hange can also move by itself, stranding the verbal particle op, thereby generating (4b). I will now briefly discuss how Bhatt and Schmidt arrived at the conclusion that am is the head of AspP. Given that am is a contraction of the preposition an and the dative neuter singular definite article dem, the analysis that immediately suggests itself is that am-progressive constructions are actually preposition phrases in which the object of the preposition is the infinitive, which has been nominalized. However, even though this may be the most obvious possible analysis of am-progressives, there is very convincing evidence that no PP is involved. As shown in Bhatt and Schmidt (1993), am-phrases do not behave at all like PPs with nominalized infinitives. There is in fact a kind of progressive construction in German that really does involve a PP with a nominalized infinitive, and the am-phrase behaves quite differently from it. I am referring to constructions with the preposition bei ‘by’ or ‘at’. These constructions are similar to am-constructions because they involve beim, a contraction of the preposition bei and the dative singular neuter form of the definite article dem: (7) Sie sind beim Essen. they are by the eat-INF ‘They are eating.’ Progressives with am behave differently from bei-phrases in important ways. Bhatt and Schmidt reason as follows. If am-phrases were PPs with a nominalized infinitive, the nominalized infinitive should be modifiable. This is however not the case 3 : (8) a. Er ist am vorlesen. he is PROG lecture-INF ‘He is lecturing.’ b. *Er ist am lauten vorlesen. he is PROG loud lecture-INF ‘He is lecturing loud.’ The word lauten in (8b) is not an adverb but an inflected adjective in the dative case. Since the infinitive cannot be modified by an adjective, it has clearly not been nominalized and is not the object of a preposition. If one uses a beim-construction instead of an am-construction, the results are quite different, because the infinitive can be modified by an inflected adjective, indicating that it is the nominalized object of a preposition: (9) a. Er ist beim Vorlesen. he is by the lecture-INF ‘He is lecturing.’ 3 Examples (7), (8), (9), (11) and (12) are from Bhatt and Schmidt (1993), 79-80. On variation in the positioning of the progressive marker am in non-standard German 9 b. Er ist beim lauten Vorlesen. he is by the loud lecture-INF ‘He is lecturing loud.’ What these data clearly show is that in bei-phrases the infinitive has truly been nominalized while in am-phrases the infinitive is still a verb. Bhatt and Schmidt show further evidence of this. In German, the complements of nouns, including nouns derived from verbs, are in the genitive case: (10) a. Die Entdeckung Amerikas the discovery America-GEN ‘the discovery of America’ b. Die Krönung des Prinzen the crowning DEF.GEN prince-GEN ‘the coronation of the prince’ In the case of bei-phrases the infinitive always takes a genitive complement, like a noun, never an accusative one. This clearly indicates that the infinitive has been nominalized and is the object of a preposition: (11) a. Er ist beim Vorlesen der Bibel he is by the read (aloud)-INF DEF.GEN bible.GEN ‘He is reading (aloud) the bible.’ b. *Er ist beim Vorlesen die Bibel. he is by the read-INF DEF.ACC bible.ACC ‘He is reading (aloud) the bible.’ In the case of am-phrases, however, the infinitive takes an accusative object like a verb: (12) a. Er ist die Bibel am vorlesen. he is DEF.ACC bible.ACC PROG read-INF ‘He is reading (aloud) the bible.’ b. *Er ist der Bibel am vorlesen. he is DEF.GEN bible.GEN PROG read (aloud)-INF ‘He is reading (aloud) the bible.’ The use of the genitive in bei-phrases and the accusative in am-phrases strongly suggests that bei-phrases are PPs while am-phrases are VPs. Further evidence of this is that in bei-constructions the complement or object of the infinitive follows that infinitive, just as the complement of a deverbal noun follows that deverbal noun, as exemplified in (10), while in the case of am-constructions the complement or object of the infinitive precedes it, just like the object in any normal VP. The only possible conclusion is that am-phrases are VPs, not PPs with a nominalized infinitive. The word am in these constructions is a progressive marker that most probably has resulted from the reanalysis 10 R o b e r t C i r i l l o or grammaticalization of the contraction am from an dem, and the infinitive is just that – a non-nominalized infinitive. Continuing with Bhatt and Schmidt’s analysis, if am-phrases are VPs, what is am exactly? It can only be one of two things. It can be inflection, that is, a progressive prefix comparable to the progressive suffix -ing found on present participles in English, or it can be the head of an Aspect Phrase. Bhatt and Schmidt convincingly rule out the possibility that it is inflection. If it were preverbal inflection, one would expect it to always be affixed to V like other forms of preverbal inflection such as ge- (the perfect aspect marker) and zu-(the infinitival marker related to the English to). This is however not the case. Compare the following examples from Standard German, in which the prefixes ge- and zu- follow the verbal particle auf ‘up’ while am precedes it 4 . (The examples with am are highly colloquial or regional.) (13) a. Der Mond ist im Begriff, auf zu gehen . the moon is in the concept up to go-INF ‘The moon is about to rise.’ b. Der Mond ist auf-ge- gangen. the moon is up PERF gone-PAST PARTICIPLE ‘The moon has risen. c. *Der Mond ist auf am gehen. (Colloquial Standard German) the moon is up PROG go-INF ‘The moon is rising.’ d. Der Mond ist am auf-gehen. (Colloquial Standard German) the moon is PROG up go-INF ‘The moon is rising.’ The same holds true for items other than verbal particles. For example, if a word such as Rad ‘bicycle’ is incorporated into a verb, the same pattern can be observed. This is shown in the following examples. (Again, the examples with am are highly colloquial or regional.) (14) a. Ursula versucht, rad- zu- fahren. Ursula tries bicycle to ride-INF ‘Ursula tries to ride a bicycle.’ b. Ursula ist rad- ge- fahren. Ursula is bicycle PERF ridden-PAST PARTICIPLE ‘Ursula has ridden a bicycle.’ c. *Ursula ist Rad am fahren. Ursula is bicycle PROG ride-INF ‘Ursula is riding a bicycle.’ 4 Examples (13)-(16) are the author’s, since Bhatt and Schmidt (1993) do not provide complete sentences to illustrate this point. On variation in the positioning of the progressive marker am in non-standard German 11 d. Ursula ist am radfahren. Ursula is PROG bicycle ride-INF ‘Ursula is riding a bicycle.’ The reader will recall that, as shown in example (4), Kölsch differs from Standard German when am co-occurs with a particle or an item that has been incorporated into the verb in that am can immediately precede the verb like the prefixes ge- and zu-. The following examples from Kölsch illustrate this again: (15) a. D’r Mond es op am jonn. the moon is up PROG go-INF ‘The moon is rising.’ b. Et Ulla es Rädche am fahre. the Ulla is bicycle PROG ride-INF ‘Ulla is riding a bicycle.’ However, as also pointed out in example (4), Kölsch also allows the word order found in colloquial Standard German: (16) a. D’r Mond es am opjonn. the moon is PROG up go-INF ‘The moon is rising.’ b. Et Ulla es am rädchefahre. the Ulla is PROG bicycle ride-INF ‘Ursula is riding a bicycle.’ This alternative word order is unthinkable for the prefixes ge-and zu-, in Kölsch as well as in Standard German. The logical conclusion is that am is not a verbal prefix and that the word order in (15) is derived by stranding the verbal particle or complement when the verb moves from V to Asp as discussed at the very beginning of this Section. The fact that this kind of stranding is available in Kölsch (and, by the way, in Dutch), but not in Standard German, can be attributed to micro-parametric variation. Bhatt and Schmidt justifiably conclude that am is not inflection and is therefore a head. Based on their very convincing arguments, I will assume that am is the head of a head-final AspP, that verbs undergo head-to-head movement and right adjoin to Asp, and that depending on the setting of micro-parameters a verbal particle or incorporatum may be stranded. In the next section we will examine a challenge to Bhatt and Schmidt posed by another West Germanic language that employs the am-progressive. 3. The am-progressive in Pennsylvania Dutch Pennsylvania Dutch is a variety of German, not Dutch 5 . It is the language of the Amish and Mennonites in the United States and is spoken by over 300,000 people, 5 The words Deutsch and Deitsch sounded like Dutch to the English-speaking Pennsylvanians, so the term Pennsylvania Dutch became standard, even for many Pennsylvania Dutch speakers. 12 R o b e r t C i r i l l o mainly in the states of Pennsylvania, Ohio and Indiana. Its origins are in a language that was spoken in the southern part of the West Middle German linguistic region and the northern part of the Upper German linguistic region in the 17 th and 18 th centuries, when the emigration of the Amish from Germany to America began. Like Kölsch, Pennsylvania Dutch also makes use of the am-progressive, but the position of am in VP is very different in the two languages. As we have seen, in Kölsch am immediately precedes the verb. If the verb is a particle verb or has incorporated a complement, am may precede the particle or incorporatum, but it will always follow objects and adverbial phrases. This is clearly illustrated in (4a), repeated here: (4) a. D’r Pitter es de Wäsch em Jade am ophange the Pitter is the laundry in the garden PROG up hang-INF ‘Peter is hanging up the laundry in the garden.’ In Pennsylvania Dutch, however, am precedes the entire VP, including direct objects, indirect objects, PPs and adverbial phrases. Because the Amish and Mennonites are very religious, the bible is a convenient source of data. The following sentences, taken from Hans-Bianchi (2013), are from the New Testament: (17) a. Da Judas voah am eena da vayk veisa. the Judas was PROg them the way show-INF ‘Judas was showing them the way.’ (Gospel of Luke 22: 47) b. Dei yingah sinn am ebbes du your disciples are PROG something do-INF ‘Your disciples are doing something.’ (Gospel of Matthew 12: 2) c. Avvah di anra yingah sinn nei kumma mitt em boat am but the other disciples are in come with a boat PROG ’s nett foll fish hinnich eena nohch zeeya the net full fish behind them still pull-INF ‘But the other disciples have come in in a boat still dragging the net full of fish behind them.’ (Gospel of John 21: 8) Examples from actual conversations are also available 6 : (18) a. Ich heb g’herd, es diah am en Haas ufdu sinn. I have heard that they PROG a house up-do-INF are ‘I have heard that they are building a house.’ b. Diah sinn am in Nappanee en Haas ufdu. they are PROG in Nappanee a house up-do-INF ‘They are building a house in Nappanee.’ 6 From conversations between the author and residents of the towns of Shipshewana and Bremen in Indiana. On variation in the positioning of the progressive marker am in non-standard German 13 c. Diah sinn am drous de schire austreicha. they are PROG outside the barn paint-INF ‘They are outside painting the barn.’ d. Samuel iss am in der Wuhnschtubb seim Geschwischderkind 7 Samuel is PROG in the living room his cousin en Brief schreiba. a letter write-INF ‘Samuel is in the living room writing his cousin a letter.’ In these examples we see am appearing before an indirect object, a direct object, a PP and an adverb. One can hardly propose that it is a verbal prefix. It can also not be a preposition, for the same reasons that it could not be a preposition in Kölsch. (If it were a preposition, for example, its complement infinitives would have to be nominalizations, but it if the infinitive complements were nominalizations their DP direct object complements would be in the genitive case. All of the direct objects in (18) are in the accusative.) The only possibility is that am is the head of an Aspect Phrase or Progressive Phrase. This raises the question of how the correct word order, with am preceding an entire VP, can be generated within a head-final verbal domain. There are two possibilities. The first possibility is that the entire VP, including objects and adjuncts, moves to the right (which is possible in a head-final environment) and right attaches to the progressive head am. A glance at (6a) will help the reader visualize how this would work. I view this as an impossible solution, for at least two reasons. First of all, there is the question of what would motivate the movement of an entire VP, including adjuncts. One can of course always postulate some kind of linearization feature on am that triggers movement, a sort of EPP feature, but I would view this as an ad hoc solution – a solution with no independent motivation. A more serious problem with a movement solution is the question of a landing site for VP. VP-movement and remnant VP-movement are of course nothing new in the Germanic languages, but they invariably involve movement to a specifier position and they are not obligatory. In the case of Pennsylvania Dutch, the positioning of am before the entire VP is obligatory. Furthermore, assuming, following Kayne (1994) and Haider (2003, 2010), that specifier positions, particularly in the Germanic languages, are always to the left of the branching node, the kind of movement that we are observing in Pennsylvania Dutch could not possibly be to a specifier position. Rather, it would have to be some kind of adjunct position. An obligatory movement to an adjunct position that is not at all clearly defined would be a very undesirable innovation. This leaves one solution: In Pennsylvania Dutch, just as in Kölsch, am is the head of an AspP or ProgP, except that in Pennsylvania Dutch ProgP is head-initial. This solution is not at all unattractive, since it allows us to explain an instance of variation in terms of a parameter 7 In Standard German the word Geschwister means ‘sibling’. The equivalent in Pennsylvania Dutch can mean ‘cousin’. 14 R o b e r t C i r i l l o – head directionality. There is, however, one potential obstacle to this approach, namely, the Final-Over-Final Constraint. The Final-Over-Final Constraint, or FOFC, was first formulated by Holmberg (2000) and further developed by Biberauer et al. (2007, 2010). It is worded as follows: “A head-initial category cannot be the immediate structural complement of a head-final category within the same extended projection”. This constraint contains the following three statements: (i) If all phrases in an extended projection are either head-initial, as in English, or head-final, as in Japanese, there are no grammaticality issues. (ii) If the head of a head-initial phrase selects a head-final phrase as its complement, there are likewise no syntactic problems. (iii) The problem arises if the head of a head-final phrase selects as its complement a head-initial phrase. As is well-illustrated in (18a), Pennsylvania Dutch would be in flagrant violation of the FOFC because the phrase headed by the progressive auxiliary sinn ‘are’ is a head- final phrase that dominates a head-initial phrase headed by am. This is demonstrated in the tree diagram in (19). Biberauer et al. (2010) argue that the FOFC can be derived from the Linear Correspondence Axiom (LCA) presented in Kayne (1994) and Relativized Minimality, first presented by Rizzi (1990), and they do offer a lot of evidence that the FOFC exists. I have argued, however, that the only plausible explanation for the positioning of am in progressive constructions in Pennsylvania Dutch is that am is the head of a head-initial phrase that can, in violation of the FOFC, be selected by the head of a head-final Auxiliary Phrase. The question is whether my approach should be discarded just because it is in apparent conflict with the FOFC. I think not, for four reasons. First of all, regardless of the FOFC, analyzing am as the head of a head-initial Progressive Phrase (even if it can be dominated by a head-final AuxP) is the only plausible explanation for the positioning of the progressive marker in Pennsylvania Dutch. All other alternatives – treating am-phrases as PPs, or treating am as a verbal prefix, or treating am-phrases as head-final – are too problematic, as we have just seen. Secondly, the FOFC is based on the assumption in Kayne (1994) that all phrases in all languages are underlyingly head- initial and that head-final word order is derived by movement. This is by no means universally accepted. Bhatt and Schmidt (1993) are not the only ones who believe that phrases can be head-final. They are joined by many others, such as den Besten (1983), Roberts (1997), Baker (2001), Grewendorf (2002), Miyagawa and Arikawa (2007), Haider (2010), and Salzmann (2013), to name but a few. On variation in the positioning of the progressive marker am in non-standard German 15 (19) AuxP 3 Aux´ 3 AspP Aux 3 sinn Asp ́ (are) 3 Asp VP am 3 (PROG) V´ 3 DP V en Haas ufdu (a house) (build) Thirdly, whereas the FOFC certainly seems to represent a trend in language, it is difficult to argue that it is an inviolable principle, for the simple reason that there are exceptions to it. Biberauer et al. (2010) offer some examples, a few of which we will now look at. The first one is from West Flemish: (20) ... da Valère willen dienen boek lezen eet 8 that Valère want-INF that book read-INF has ‘... that Valère has wanted to read that book’ In this example, willen dienen boek lezen represents a head-initial Modal Phrase headed by willen ‘want’ that is immediately dominated by the perfect auxiliary eet, which heads a head-final phase. Biberauer et al. (2010) suggest that this apparent violation of the FOFC is due to the special status of eet, which is perhaps something other than a perfect auxiliary. This is highly unlikely, however. A sentence needs a finite verb, and the only finite verb in this example is the perfect auxiliary eet. Biberauer et al. (2010) present other exceptions to the FOFC which they attempt to explain away in a similar way. The following is from Afrikaans: (21) ... dat hy die boek loop koop het that he the book walk buy has ‘...that he went to buy the book’ Here we see a head-final Perfect Phrase headed by het dominating a head-initial phrase headed by loop. Biberauer et al. (2010) derive this sentence from a base-structure 8 This example contains an infinitivus pro participio or IPP construction, in which the infinitive of willen is used as the complement of the perfect auxiliary instead of a past participle. The IPP is very common in the continental West Germanic languages. 16 R o b e r t C i r i l l o in which all phrases are head-initial, as illustrated in (22). In this model, the direct object die boek has to be scrambled to a higher position. Next, the two infinitives loop and koop have to be raised above the perfect auxiliary het. Normally, under a “roll-up” operation, koop would raise to loop and left-adjoin to it, and the two would then raise and left-adjoin to het. However, this would produce the undesired word order *koop loop het. In order to get around this, Biberauer et al. (2010) must generate loop koop as a unit meaning ‘go buy’, which is not implausible. Be that as it may, there is a lot of work involved in this derivation, and, ultimately, in the surface word order, the sentence in (21) contains a head-initial phrase that is dominated by a head-final phrase, and this is potentially a violation of the FOFC. (22) PerfP 3 Perf´ 3 Perf VP het 3 V´ 3 V VP loop 3 V ́ 3 V DP koop die boek Another example from Afrikaans of an apparent exception to the FOFC presented by Biberauer et al. (2010) is the following, which they do not explain away but leave for future research: 9 (23) ... dat hy haar hoor kom het that he her hear-INF come-INF has ‘... that he has heard her come’ In this sentence, the head-initial phrase headed by hoor is dominated by a head- final Perfect Phrase headed by het. To clearly illustrate this, I offer the equivalent sentence in Dutch, in which there is no violation of the FOFC because the Perfect Phrase is head-initial: 9 Biberauer et al. (2010) point out that this example contains an IPP structure and they suggest that the apparent violation of the FOFC may be able to be explained if more is learned about IPP. On variation in the positioning of the progressive marker am in non-standard German 17 (24) … dat hij haar heeft horen komen. that he her has hear-INF come-INF ‘... that he has heard her come’ This Dutch example also contains an IPP (see footnote 8), since the complement of the perfect auxiliary heeft, which is horen, is an infinitive instead of a past participle. Incidentally, the most elegant way to derive this Dutch sentence is to posit underlying head-final verbal phrases. This would involve a simple roll-up and would avoid both object raising and the base-generation of horen komen as a unit. The reader is referred to the tree diagram in (25). The verb komen raises to horen and right-adjoins to it, and the combination horen komen is raised and right-adjoined to heeft. (25) PerfP 3 Perf´ 3 VP Perf 3 heeft V´ 3 VP V 3 horen haar V ́ 3 V komen The fourth and most important reason that I want to discuss for not abandoning my approach just because of an apparent violation of the FOFC is that a head-initial am-phrase dominated by a head-final AuxP may not be a violation of the FOFC after all. I will elaborate. Biberauer et al. (2010) base their arguments regarding the verbal domain mainly on constructions involving Aux, V and O. They offer cross-linguistic evidence of the following gap in the paradigm of word order: (26) a. Aux V O (purely head-initial) b. O V Aux (purely head final) c. Aux O V (head-initial Aux Phase dominating head-final VP) d. *V O Aux (head-final Aux Phrase dominating head-initial VP) They do not include particles in their discussion of the Germanic languages. They do say, however, that the FOFC applies within an extended projection, and one would assume that an am-phrase would be the extension of a verbal projection. On the other hand, they also point out that outside the Indo-European family there are some exceptional examples involving aspect particles. These are instances in which a head-final Aspect Phrase 18 R o b e r t C i r i l l o headed by an uninflected aspect marker/particle can felicitously dominate a head-initial VP. They present the following example from Bwe-Karen, a Sino-Tibetan language spoken in Burma: (27) yә - ca dɛyo lɔ 1SG-see picture ASP ‘I am looking at a picture.’ The German am could also be considered an uninflected aspectual particle. It is also important to point out that Biberauer et al. (2010) argue that a categorial difference between two projections may exempt a clause from compliance with the FOFC. An example is the following one from German, in which a head-final VP dominates a head- initial PP: (28) Johann ist nach Berlin gefahren. Johann is to Berlin driven ‘John has driven to Berlin.’ Thus, it may be that Pennsylvania Dutch does not pose an exception to the FOFC at all because the progressive marker/particle am is not of the same category as a verb. This idea becomes especially interesting if one considers that historically am was derived from a preposition and might have retained a certain residual status as such. Bhatt and Schmidt (1993) in fact offer evidence for the residual prepositional status of am. They point out the following discrepancy in colloquial Standard German: (29) a. Der Pilot ist den Airbus am fliegen. the pilot is the Airbus PROG fly-INF ‘The pilot is flying the Airbus.’ b. *Der Pilot ist den Airbus nach Wahn am fliegen. the pilot is the Airbus to Wahn PROG fly-INF ‘The pilot is flying the Airbus to Wahn.’ The explanation that they propose, very briefly, is as follows: A verb like fliegen ‘fly’ has a locative feature in its -grid and therefore assigns a locative -role. The canonical means for assigning a locative -role in the Germanic languages is of course the selection of a PP. The locative feature in the -grid of fliegen is oversaturated in (29b) because the locative PP nach Wahn ‘to Wahn’ co-occurs with the progressive marker am, which, due to its homophonous relationship with the prepositional contraction am, has a kind of residual status of preposition and can therefore trigger a second locative -role assignment by the main verb, in violation of the -Criterion. Note that this restriction applies to Standard German but not to Kölsch, indicating micro-parametric variation. Also, in my opinion, a lot of speakers find (29b) just as acceptable as (29a). Nonetheless, I do agree with Bhatt and Schmidt that the discrepancy between (29a) and On variation in the positioning of the progressive marker am in non-standard German 19 (29b) for some speakers of colloquial Standard German could be an indication that am has a residual prepositional status. Based on all the above discussion, my conclusion is that in Pennsylvania Dutch am-phrases are head-initial even though they can be dominated by a head-final AuxP. 4. Conclusions In Section 1 the phenomenon of the am-progressive in colloquial Standard German and Kölsch was introduced. In Section 2 the analysis of the am-progressive in Bhatt and Schmidt (1993) was presented. These authors argue convincingly that am is neither a preposition nor a verbal prefix and must therefore be the head of a head-final Aspect Phrase. In Section 3 the am-progressive in Pennsylvania Dutch was introduced, and I argued that am in Pennsylvania Dutch, like its equivalent in Kölsch, cannot be a preposition or a verbal prefix and must therefore be the head of an Aspect Phrase. Unlike am-phrases in Kölsch, however, am-phrases in Pennsylvania Dutch seem to be head- initial rather than head-final. If they were head-final their complement VP would not only have to move rightward but it would have no landing site. (A VP would be expected to move to a specifier position, but specifiers are to the left of the branching node, not to the right). The only obstacle to this approach is that it is in potential conflict with the FOFC because it allows a head-initial am-phrase to be dominated by a head-final AuxP. I gave four reasons why I was not concerned with this potential conflict: (i) Regardless of any conflict with the FOFC, there is no other analysis of the am-progressive in Pennsylvania Dutch that makes any sense. (ii) The FOFC is based on the assumption in Kayne (1994) that all phrases in all languages are underlyingly head-initial. This is anything but universally accepted. (iii) There are exceptions to the FOFC, which implies that whereas it indicates a clear trend in language it is not an inviolable principle. (iv) It is possible that the present analysis of am-progressives in Pennsylvania Dutch is not incompatible with the FOFC. The initiators of the FOFC recognize that uninflected particles (of which am could be considered to be one) are sometimes involved in situations in which a head-final phrase dominates a head-initial one. Furthermore, one might be able to argue that am is categorially distinct enough from verbs that projections involving the am-progressive could be exempted from the FOFC. This idea is particularly interesting given the fact that am originated as a preposition and may well have a certain residual status as such. References Baker, M. 2001. The Atoms of Language. New York: BasicBooks. den Besten, H. 1989. Studies in West Germanic Syntax. Amsterdam, Atlanta: Rodopi. Bhatt, C., Schmidt, C.M. 1993. Die am + Infinitiv-Konstruktion im Kölnischen und im umgangssprachlichen Standarddeutschen als Aspektphrasen. In W. Abraham and J. Bayer (eds), Dialektsyntax (Linguistische Berichte, Sonderheft 5), 71-98. Opladen: Westdeutscher Verlag. 20 R o b e r t C i r i l l o Biberauer, T., Holmberg, A., and Roberts, I. 2010. A syntactic universal and its consequences. Ms., University of Cambridge/Stellenbosch University, Newcastle University and University of Cambridge. Biberauer, T., Holmberg, A., and Roberts, I. 2007. Disharmonic word-order systems and the Final-over-Final- Constraint (FOFC). In A. Bisetto and F. Barbieri (eds.), Proceedings of XXXIII Incontro di Grammatica Generativa, 86-105. Università di Bologna. http://amsacta.cib.unibo.it/archive/0000 2397/01/PROCEEDINGS_IGG33.pdf. Grewendorf, G. 2002. Minimalistische Syntax. Tübingen and Basel: A. Franke Verlag. Haider, H. 2003. V-clustering and clause union-causes and effects. In P. Seuren and G. Kempen (eds), Verb Constructions in German and Dutch, 91-126. Amsterdam/Philadelphia: John Benjamins. Haider, H.2010. German Syntax. Cambridge, New York: Cambridge University Press. Hans-Bianchi, B. 2013. Die geborgte Grammatik: zum Phänomen kontaktinduzierter Grammatikalisierung am Beispiel des Pennsylvania Deutsh. In B. Hans-Bianchi, L. Zenobi, C. Miglio, D. Pirazzini, I. Vogt (eds.), Fremdes wahrnehmen, aufnehmen, annehmen – Studien zur deutschen Sprache und Kultur in Kontaktsituationen, 195-214. Frankfurt am Main: Peter Lang. Holmberg, A. 2000. Deriving OV order in Finnish. In Peter Svenonius (ed.), The Derivation of VO and OV, 123-152.Amsterdam/Philadelphia: John Benjamins, Kayne, R. S. 1994. The Antisymmetry of Syntax. Cambridge, MA: The MIT Press. Miyagawa, S., Arikawa, K. 2007. Locality in syntax and floating numeral quantifiers. Linguistic Inquiry 38: 645-670. Rizzi, L.1990. Relativized Minimality. Cambridge, MA: The MIT Press. Roberts, I. 1997. Comparative Syntax. London: Arnold. Salzmann, M. 2013. Rule ordering in verb cluster formation: On the extraposition paradox and the placement of the infinitival particle te/zu. In F. Heck and A. Assmann (eds.), Rule Interaction in Grammar, 65-121. Leipzig: Universität Leipzig.