This essay will attempt to demonstrate why coordination allows the use of bare definites in Italian in cases in which they would be otherwise ungrammatical. This has always been a mysterious and unresolved ‘phenomenon’, which has caused embarrassment to linguists.
This essay is structured as follow: section 1 will individuate the difference between the distribution of bare definites in Greek and the one in Italian; in section 2 some examples will be given to show how coordination can allow some syntactical constructions that are not otherwise grammatical and different kinds of readings of coordination will be presented in Greek and Italian; in section 3 Longobardi’s analysis of N-to-D movement will be summarized; in section 4 the critics made by Heycock and Zamparelli to Longobardi’s analysis will be enumerated; in section 5 the ‘amended’ account given by Heycock and Zamparelli will be presented; in section 6 the critics made by Roodenburg to H&Z account will be examined; in section 7 an account of ‘and’ to D movement will be given; in section 8 some further investigations in prosody are suggested; in conclusion an evaluation of the whole picture will be given.
Comparison between Italian and Greek distribution of bare nouns.
Firstly, this essay will draw a comparison between the use of bare definites in Italian and the one in Greek. As Alexopoulou, Folli and Tsoulas (2013) say in their paper ‘Bare number’, the distribution of bare nouns is much more liberal in Greek than in Italian. Indeed, ‘it is well known that bare nouns are licit in very restricted positions in Italian, roughly in governed positions (objects) and focalised positions’ (Alexopoulou, Folli and Tsoulas, 2013 (Benincà 1980, Longobardi 1994, Chierchia 1998)). On the other hand, ‘in Greek bare nouns can be found both in preverbal and post-verbal subject positions as well as in object position’ (Alexopoulou, Folli and Tsoulas, 2013). Therefore, the main difference seems to be the use of bare nouns in preverbal subject position, which is allowed in Greek, but not in Italian.
1.1 Distribution of bare nouns in Italian (Longobardi 1994).
Empy determiners may occur at S-Structure in Italian only under the following conditions:
– They are restricted to plural or mass head nouns like several other determiners.
(1) a. # Mangio patata. (Singular count noun)
‘I eat/am eating potato’.
- Mangio patate. (Plural count noun)
‘I eat/am eating potatoes’.
- Bevo sempre vino. (Mass noun)
I drink always wine.
‘I always drink wine’.
– They are subject to a lexical government requirement like other empty heads.
(2) a. # Acqua viene giu’ dalle colline. (Bare noun in pre-verbal position)
‘Water comes down from the hills’.
- Viene giu’ acqua dalle colline. (Bare noun in post-verbal position)
Comes down water from the hills.
‘Water comes down from the hills’.
– They receive an indefinite interpretation corresponding to an existential quantifier
(3) a. # Amo foreste tropicali. (Generic reading)
I love tropical forests.
‘I love tropical forests’.
- Ho sempre studiato foreste tropicali. (Existential reading)
I have always studied forests tropical.
‘I have always studied tropical forests’.
- Ho sempre studiato le foreste tropicali. (Generic reading)
I have always studied the forests tropical.
‘I have always studied the tropical forests’.
1.2 Distribution of bare nouns in Greek
Unlike in Italian, in Greek bare nouns can appear not only as plurals and mass nouns but also as singular count nouns:
(4) Kreiazomai kapelo
I need hat
Unlike in Italian, in Greek bare nouns can appear, not only in ‘lexical governed’ position, but also in pre-verbal position:
(5) Dadilotes pirpolisan magazia ke aftokinita stus dromus yiro apo to
Demonstrators set-on-fire shops and cars in-the streets around from the Polithenio
However, as pointed out originally by Roussou and Tsimpli (1994) (Alexopoulou, Folli and Tsoulas, 2013), even in Greek, as well as in Italian, bare plural nouns in subject position cannot have the generic/kind reading, but only the existential one.
(6) a. ta/ * (theta) skilia ine katikidhia zoa. (Generic reading)
the/ * (theta) dogs are domestic animals.
- Dinosavri ehun eksafanisti.
‘Dinosaurs have disappeared’. (Bare nominal ungrammatical under the generic/kind reading, it is possible only under the existential reading).
- Reference to Kinds in Italian and in Greek
This paragraph will be devoted to analyse the similarities in the distribution of bare nouns with reference to kinds in Italian and in Greek.
The first similarity is the fact that both in Italian and in Greek a single event verb ‘found’ or a durative verb ‘study’  (stage-level predicates) favor an existential reading of the object and therefore it makes the sentence grammatical, as in the example (7a-b), whereas the permanent state verbs ‘love’ and ‘hate’ (individual-level predicates) force the generic one, incompatible with bare nouns, as in (8 a-b) (vd Carlson)
(7) a. Ho trovato arance fresche
I found fresh oranges.
- O Kostas meleta karkinous
The Kostas study-3SG cancers
‘Kostas studies cancers’
(8) a. *Amo arance fresche.
I love fresh oranges.
- * O Kostas misei karkinous
The Kostas hates-3SG cancers
‘Kostas hates cancers’.
The second similarity is that between the individual -level -predicates the ‘eventive’ ones are grammatical, while the more ‘stative’ ones are not grammatical. For example (9 a-b)
(9) a. ?? Stati di grandi dimensioni sono prosperi. (stative)
?? Megala se megethos krati einai euimera
‘States of large size are prosperous’.
- Stati di grandi dimensioni sono pericolosi (eventive)
megala se megethos krati einai epikindyna
States of large size are dangerous.
Moreover, (10 a) is ungrammatical, because, not only the noun is not preceded by the article, as in (10 c), but also there is no ‘characterizing environment’, such as the adverbial phrase ‘in the Saronic gulf’ in (10 b):
(10) a. * Falenes spanizoun
- Falenes spanizoun sto Saroniko
‘Whales are rare in the Saronic gulf’.
- oi falenes spanizoun
the whales-NOM are-rare
‘whales are rare’
In conclusion, in Italian and in Greek, predicates used in a ‘characterizing environment’ are the most grammatical for bare nouns both in subject and in object position. (?)
Interestingly, both in Italian and Greek, coordination can make grammatical bare nouns that are otherwise unavailable.
In Italian a bare noun, if in coordination, can overcome all the restrictions analysed so far
i.e. it can be a singular count noun, in pre-verbal position and with the generic/kind reading, such as in (11)
(11) Cane e gatto sono sempre nemici.
‘Cat and dog are always enemies’.
In Greek bare plurals in subject position, if coordinated, can be used to denote kinds, for example (12)
(12) Gates kai schyloi den ta pane kala.
‘Cats and dogs don’t get along’
Therefore coordination, making bare nouns grammatical, seems sufficient to entirely subvert/undermine the picture of distribution of bare nouns in Italian and Greek outlined so far. This fact has always been somewhat an embarrassment, because it is not clear at all why coordination is enough to change the grammaticality of certain syntactical structures. Many linguists have tried to give a satisfactory account of coordination which could explain this ‘anomaly’ but they still seem incomplete.
2.1 Different Kinds of Coordination
In order to give a complete account it is also important to analyse the different kinds of coordination between bare definites. The first important distinction is the one between the ‘joint reading’ and the ‘split reading’. Moreover, the two readings differ in singular and in plural.
The joint reading of coordination is when two coordinated names refer to the same entity. Indeed (13) is an example of joint reading in the singular and (14) is an example of joint reading in the plural.
(13) [DP My [NP friend and colleague]] is writing a paper.
(14) [My two [friends and colleagues]] wrote their paper together. (According to the joint reading there are two people, each of whom is both a friend and a colleague at the same time)
On the other hand, the split reading is when the two coordinated names refer to two different entities. (15) is an example of split reading in the singular and (16) is an example of split reading in the plural.
(15) [My grandfather and great grandfather] were both sailors.
(16) [His friends and colleagues] all came to the party.
- Longobardi’s Account of N-to-D movement.
According to Longobardi, in Romance languages, such as Italian, the referential feature of the determiner position, D, is ‘strong’; that is, visible systematic association of referential items with D is necessary. In other languages, such as English, the referential properties of D are ‘weak’ i.e. referential reading may affect nominal items not overtly associated with D.
That is why, unlike in English, ‘in Romance, object-referring nouns in argument function are always necessarily introduced by a phonetically expanded D node: either they occur after a visible determiner or are themselves moved to D, giving rise to typical patterns like (17 a-c)
(17) a. # Antica Roma (fu distrutta dai barbari)
‘Ancient Rome (was destroyed by the barbarians)’
- Roma antica (fu distrutta dai barbari)
‘Rome ancient (was destroyed by the barbarians)’
- L’antica Roma (fu distrutta dai barbari)
‘The ancient Rome (was destroyed by the barbarians)’
(17 a) is ungrammatical because, although there is no article, N has not raised to D; on the other hand (17 b) is grammatical because, since there is no article, N has raised to D; (17 c) is grammatical because, since there is the article, N does not need to raise to D position.
In conclusion, the N-initial order in the articleless examples is obligatory because argument nominals need to be introduced by a D position, which cannot be left empty at S-Structure.
However, there is an important issue in the N-to- D movement theory. Although this theory shows that proper names and definites behave in the same way (?) in Romance, it does not take into account the fact that they differ in terms of rigidity i.e. proper names, such as ‘Anselm’ in (21 a), designate the same ‘object’ in all possible worlds, whereas this is not the case for definites such as ‘the discoverer of the ontological proof’ in (21 b). Indeed, unlike for the proper name ‘Anselm’, it is easy to imagine a world in which the definite ‘the discoverer of the ontological proof’ designates a different ‘object’.
(21) a. Anselm was born in Aosta and became archbishop of Canterbury.
- The discoverer of the ontological proof was born in Aosta and became archbishop of Canterbury.
3.2 Relation between the syntax of PNs and the interpretations of BNs.
According to Longobardi, in English Bare Nouns (BNs), as in (18 a), and Proper Nouns (PNs), as in (18 b), have the same surface structure, which is [e (Adj) N], where ‘e’ symbolizes the lack of overt determiner.
(18) a. [e [Adj Young] N ladies] are often spoilt.
- [e [Adj Big] N Lucy] is nice.
On the other hand, in Italian, BNs and Ps do not have the same surface structure, because BNs have the structure [e (Adj) N], as in (19), while PNs have either the structure [Art (Adj) N], as in (20 a) or the structure [N (Adj) t], as in (20 b).
(19) [e [Adj Grossi] N cani] giocavano in cortile.
Big dogs were playing in the court.
(20) a. [Art L’[ Adj antica] N Roma] (fu distrutta dai barbari)
‘The ancient Rome (was destroyed by the barbarians)’
- [[N Roma] Adj antica] t] (fu distrutta dai barbari)
‘Rome ancient (was destroyed by the barbarians)’
Moreover, BNs can be referential (instantiate kind-denoting constants) iff they have the same surface structure as PNs, as in the case of English. On the other hand, in Italian BNs do not have the same surface structure as PNs, therefore BNs in Italian cannot be referential (instantiate kind-denoting constants). This explains why the generic/kind reading, which is the one that instantiates kind-denoting constants, is not available for Italian bare nouns.
To sum up, Longobardi’s account covers all cases of nominals i.e. proper names, definites, object-denoting and kind-denoting indefinites, both in Romance and in Germanic languages.
In both languages proper names, definites and kind-denoting indefinites are considered ‘referential’, whereas object-denoting indefinites are considered ‘quantificational’; however, the main difference between Romance and Germanic languages is that in Romance ‘kind-denoting generics’ are only expressible through overtly definite DPs in Romance, whereas in Germanic languages, such as in English, they are expressible also through bare nouns.
Therefore, Romance bare nouns are only quantificational expressions (i.e.variables, indefinites); they thus behave like overt indefinites and unlike proper names. English bare nouns can also be referential (i.e. generic constants, kind names); they thus behave unlike overt indefinites and like proper names (cf. Carlson 1977a, b). (Longobardi)
- Critics to Longobardi’s account
In their paper ‘Coordinated Bare Definites’ H&Z (2003) have made a list of remarks to Longobardi’s theory of N-to-D movement.
They introduce the section saying that Longobardi (1994) has proposed that a definite meaning can be triggered by N-to-D rasing. This suggests that bare singular coordination could involve the coordination of two DPs, in each of which N has raised to D’.
(21) [Coord [DP goblet i … t i] and [DP spoon j …tj]]
However, this analysis has arisen some problems:
1) N-to D raising predicts that the bare singular conjuncts should not allow modification, such as (22)
(22) … [Crystal goblet and golden spoon] had to be expressly taken from the jewel room.
According to Longobardi, the premise itself of the first remark made by H&Z to his account is wrong. H&Z say that ‘Longobardi (1994) has proposed that a definite meaning can be triggered by N-to-D raising (as in the case of proper names)’. However Longobardi says that it is not a ‘definite’ meaning which can be triggered by N-to-D raising but rather that a ‘referential’ one. The main difference between ‘referential’ and ‘definite’, as it has already been said, is that the former one is rigid i.e. it designates the same objet in all the possible worlds, as in the case of proper names, while the second one is not.
2) The second problem arisen by H&Z is that Longobardi’s cases of N-to-D raising display rigidity of reference, typical of proper names, while the meaning of a bare singular coordination does not have the semantic rigidity
(23) a. Camera mia cambiera’ (semantic rigidity after N-to-D raising)
Room my will change.
‘My room will change’
- La mia camera cambiera’ (semantic non-rigidity)
The my room will change
‘My room will change’
- Coltello e cucchiaio vanno messi a destra del piatto. (semantic non-rigidity)
‘Knife and spoon need to be set on the right of the plate’.
This point is connected to the previous one in the sense that ‘camera’ in (23 a), after having raised to D position, crossing over (?) the adjective, assumes a ‘semantic rigidity’ so that (23a) means that ‘my room will change’ aspect; on the other hand ‘camera’ in (23 b) has not raised to D, because that position is already taken by the article ‘la’, and so it does not assume a ‘semantic rigidity’, therefore (23 b) means that ‘I will change room’. Even though Longobardi’s account of N-to-D movement focuses mainly on proper names, the difference between (23 a) and (23 b) is actually in line with Longobardi’s account in the sense that N-to-D raising of a common noun has triggered ‘referentiality’ and therefore ‘semantic rigidity’.
However, Longobardi does not give a full explanation of the fact that coordination of bare singular nouns, as in (23 c), allows grammaticality even without triggering ‘referentiality’.
4) An N-to-D raising analysis incorrectly predicts that coordination of a bare singular with a normal definite DP or a proper name should be possible, just like John and his friend; thus, (25 a) should be grammatical with the structure given in (25 b)
(25) a. # … fork and the spoon must be…
- … [DP fork] and [DP the spoon]] must be …
5) Coordination has no special licensing effect on bare singulars when it applies at the sentential level (27 a), or to predicate nominal without uniqueness presuppositions (27 b)
(27) a. Fork is silver-plated and bowl is enamelled.
- John is gentleman and Englishman.
6) According to H&Z, the most important remark is the fact that ‘the N-to-D raising analysis does not suggest any reason why coordination should be a necessary condition for the occurrence of bare singulars’; that is, it provides no insight into the ungrammaticality of an example like (26)
(26) # … [DP Fork] must be placed on the left.
However, this point does not stand because Longobardi does not say that coordination is a ‘necessary’ condition for the occurrence of bare singulars, he only says that coordination is a ‘sufficient’ condition to allow grammaticality of bare singulars.
In conclusion, after having listed all the ‘problems’ of Longobardi’s account stressed by H&Z and after having tried to confute some of them, this essay will present an account of bare noun coordination given by H&Z who retained Longobardi’s insight and tried to amend it.
- Heycock and Zamparelli’s account of bare noun coordination.
The solution given by Heycock and Zamparelli, as an alternative to Longobardi’s account is that bare noun coordination involves movement of a [CoordP NP Coord NP] structure to a single Spec, DP, with D empty, as illustrated schematically in (28)
(28) [DP [CoordP[NP goblet] and [NP spoon]]i [D’ De…ti]]
(29) [Coord [DP goblet i … t i] and [DP spoon j …tj]]
The main difference between this account and the one given by Longobardi is that according to Longobardi’s one, as in (29), the overall nominal structure is a coordination between two DPs and in each DP the noun moves to D position. On the other hand, according to the new account given by Heycock and Zamparelli (H & Z), as shown in (28), the overall structure is a DP and the D position is filled by the entire coordination ‘goblet and spoon’, which has raised to D.
The reason why, according to H&Z, the whole CoordP needs to raise to Spec,DP, unlike in Longobardi’s account, is because this is the only way to create a configuration of agreement that allows feature sharing between and and De. Indeed, De head is empty and in need of identification; the way to identificate De is to allow the share of the [+ Qu] feature, with which and is endowed, between and and De. However, [+ Qu] feature can be assigned to De only under predication, but when the bare noun is a singular count ([-Plur]), as in the case of (28), it cannot function as predicate nominal. Therefore pied-piping of the whole CoordP to Spec,DP is necessary in order to license De, in the case of singular bare nouns.
- Critics to H&Z account, made by Roodenburg.
- According to H&Z, French Coordinated Bare Nouns (CBNs) may have only a ‘’definite’’ reading, and do not allow an existential one. However, according to native speakers of French, the CBP in (30) is perfectly all right with an existential reading.
(30) Jour de marche’ en ville; ?? clients et curieux se promenent autour des etalages.
‘Market day in town; customers and onlookers walk about the stands’.
Therefore, (30) is a counter-evidence for the derivation made by H&Z, according to which, ‘an existential reading is obtained when the CoordP stay in its base position’.
- According to Roodenburg, H&Z’s claim that ‘the specification of NumP depends on that of the individual conjuncts’ is wrong. On the contrary he hypothesizes that all CBNs are characterized by the presence of a [+pl] feature in NumP. Indeed, if H&Z are correct in their claim that only [+pl] can be used as predicate nominal in Italian/English, the nouns in (31 b) and (32 b) must be considered [+ pl].
(31) a. He was winner
- He was (both) winner and loser.
(32) a. * Il gatto e’ mammifero.
‘The cat is mammal’.
- Il gatto e’ (al tempo stesso) vertebrato e mammifero.
‘The cat is (at the same time) vertebrate and mammal.
- According to Roodenburg, H&Z tacitly assume that [+ qu] can fulfil two different roles: (a) that of a formal licenser, or (b) that of an operator binding variable. On the contrary, Roodenburg claims that [+qu] can only be a quantificational operator and that is absent in non-quantificational structure.
- Coordination as a quantificational operator.
This last point is very important, because it stresses the fact that all the accounts considered so far – Longobardi, H&Z and Roodenburg- agree on the fact that the coordination and is a quantificational operator, which carries the [+qu] feature. This is the starting point on which to build a more complete and satisfactory account of the fact that coordination allows the grammaticality of bare nouns in Romance languages.
The core of this new account is the fact that what moves in De position is not the name itself, as claimed by Longobardi, and not even the whole CoordP, as claimed by H&Z, but the ‘and’, which is a quantificational operator, which carries the feature [+qu]. What triggers the covert movement of ‘and’ to De position is the fact that De is empty and in the need of identification. Therefore the covert movement of and-to-D licenses De.
The movement to D of the quantificational operator ‘and’ implies plurality i.e. at least two entities. Indeed, in the case of ‘the joint reading’, in which the referent of the two nouns is always the same and therefore only one entity is involved, the article is obligatory, as in (33 a), therefore there is no need for ‘and’ to raise to D. On the other hand, when there is no article the ‘and’ covertly moves to D and ‘assumes’ immediately the meaning ‘at least two’ forcing ‘the split reading’, as in (33b):
(33) a. Il collega e amico e’ un lavoratore instancabile (joint reading)
The-SING colleague and friend is a worker tireless
‘The colleague and friend is a tireless worker’.
- Collega e amico camminano insieme (grammatical only under the split reading)
‘Colleague and friend are walking together’
In (33a) the article takes the scope over the conjunction ‘and’, and gives its singularity to the verb; whereas in (33b), since there is no article, ‘and’ has the dominant scope. After having covertly risen to D ‘and’ creates plurality-because it implies ‘at least two’, thus giving its plurality to the verb and forcing the split reading. That’s why it is not possible to have a singular joint reading in case of coordination of two bare nouns, as shown by (33b).
Moreover, in Italian ‘the split reading’ in the singular, when the singular article takes scope over both the nouns and the verb is in the plural, results ungrammatical, as in (34), because there is a mismatch between the singular article and the verb in the plural, unlike in the case of (33 a) where both the article and the verb were singular
(34) * Un uomo e bambino mangiano (split reading).
A man and child are eating
However, even a sentence in which both the article and verb are plural, but the nouns are singular is not grammatical, because there is a mismatch between article and verb on one hand, and nouns on the other, such as in (35)
(35) * Degli uomo and bambino mangiano
The-plur man and child are eating
Therefore, the only possible singular split reading in Italian is given by the coordination of two singular bare nouns, as in (33 b).
(33) b. Collega e amico camminano insieme (grammatical only under the split reading)
‘Colleague and friend are walking together’
It is now interesting to see how the plural works and compare it with the singular.
(36) a. Gli amici e nemici di Gianni si trovano d’accordo su un solo punto (split reading)
The friends and enemies of Gianni are in agreement on a single point.
‘Gianni’s friends and enemies are in agreement on a single point’.
- I colleghi e amici di Gianni giocavano a biliardo (joint reading).
The colleagues and friends of Gianni were playing pool.
‘Gianni’s colleagues and friends were playing pool’.
- Colleghi e amici di Gianni giocavano a biliardo (ambiguous)
Colleagues and friends of Gianni were playing pool
‘Gianni’s colleagues and friends were playing pool’.
(36 a) can only have a ‘split reading’ for semantic reasons: the meaning of the two nouns is exclusive in the sense that someone cannot be ‘friend’ and ‘enemy’ at the same time; moreover this reading is also confirmed by the verb ‘to be in agreement’, which makes it very plausible to have two distinct factions, one of friends and one of enemies, who find a point in common. Therefore despite of the fact that there is only one article which takes scope over the two nouns, the sentence is forced to have the ‘split reading’.
On the other hand, this does not apply for (36 b), because it is much more plausible that there are some people who are ‘friends are colleagues’ at the same time than ‘friends’ and ‘enemies’ at the same time; moreover, the fact that there is only one article, which takes scope over both of the nouns, ‘induces’ the split reading.
(36 c) is ambiguous in the sense that it can have either a joint or a split reading, or even both at the same time: according to ‘the joint reading’, all the people are both ‘friends’ and ‘colleagues’ at the same time; according to ‘the split reading’ each person is either ‘a friend’ or ‘a colleague’, but not both -the disjunction is exclusive; according to ‘the mix reading’ there are some people who are either only friends or only colleagues and some people who are both ‘friends’ and ‘colleagues’ at the same time.
In conclusion, when there are two singular nouns in coordination the difference between joint and split readings changes the number of the entire sentence –singular in case of the joint reading and plural in case of the split one- ; therefore, a sentence with singular nouns in order to be grammatical needs to have a numerical match between article, nouns and verbs; on the other hand, with two coordinated plural nouns the difference between the two readings does not influence the grammaticality of the sentence, because the verb has to be plural in both cases.
To conclude, all these observations have demonstrated that ‘and’ is a quantificational operator, which carries the [+qu] feature and raises to De when this position is empty, giving plurality to the whole sentence. That’s why the singular joint reading with two singular bare nouns in coordination is impossible in Italian: there would be a mismatch between the plurality given by the arisen ‘and’ and the need of singularity of the entire sentence.
Since the quantificational operator ‘and’, after having raised to D, gives plurality to the whole sentence, it allows the use of singular bare nouns that would be ungrammatical if not in coordination. Basically, the quantificational operator ‘and’, when it coordinates two bare nouns, substitutes a plural article.
- Further Investigations in Prosody.
A different account of the reason why coordination supplies the lack of article can be
found in prosody. It can be assumed that there is ‘a prosodic range’. It follows that two nouns do not need the article when in coordination because, thanks to the coordination, they reach the ‘prosodic maximum’; on the other hand a noun without the article does not reach ‘the prosodic minimum’, therefore it is ungrammatical.
However, the immediate counterexamples of this theory would be: a) a case in which two nouns are in coordination and they have an article, such as in (37), because it would exceed ‘the prosodic maximum’; b) a case in which a proper name is not preceded by the article, because it would still be grammatical even without reaching ‘the prosodic minimum’.
For a matter a space this account will not be analysed any further, although it will be really interesting to explain the phenomenon of coordination not only from a semantic and syntactical point of view, but also from a prosodic one.
In conclusion, this essay has shown that coordination allows the use of bare nouns in cases in which they would be otherwise ungrammatical. Then it has examined some of the most relevant accounts given on the reason why this phenomenon applies. After having considered the core of these accounts and their ‘problems’, a new account has been given as a way to incorporate and ameliorate the previous ones. The main amelioration is that only ‘and’ moves to D, and not the nouns, or the whole CoordP, therefore the movement is lighter.
The language on which this essay has focused is Italian, although it will be really interesting to see whether the covert movement of ‘and’ to D applies as a remedy of the lack of D, even in other languages.
Alexopoulou T., Folli R., Tsoulas G. (2013), Bare Number.
Carlson, G. (1977a). Reference to kinds in English. Doctoral dissertation, University of Massachussetts, Amherst.
Carlson (1977). A unified analysis of English bare plurals. Linguistics and Philosophy 1: 413-457.
Chierchia, G. (1998). Reference to kinds across language. Natural language semantics, 6(4), 339-405.
Grice, H. P. (1970). Logic and conversation (pp. 41-58). Harvard Univ.
Heycock, C., & Zamparelli, R. (2003). Coordinated bare definites. Linguistic Inquiry, 34(3), 443-469.
Heycock, C., & Zamparelli, R. (2003). Friends and colleagues: Plurality, coordination, and the structure of DP. Unpublished Manuscript.
Longobardi, G. (1994). Proper names and the theory of N-Movement in Syntax and Logical Form
Longobardi, G. (2001). How Comparative is Semantics? A Unified ParametricTheory of Bare Nouns and Proper Names. Natural Language Semantics, 9(4), 335-369.
Lyons, C (1999). Definiteness. Cambridge Textbooks in Linguistics.
Roodenburg, J. (2004). French bare arguments are not extinct: The case of coordinated bare nouns. Linguistic Inquiry, 35(2), 301-313.
Winter, Y. (2001). Flexibility principles in Boolean semantics: The interpretation of coordination, plurality, and scope in natural language (Vol. 37). MIT press.
 Lexical government requirement:
 This can be joint and split at the same time. There can be people who are both a friend and a colleague at the same time and people who are only friends or only colleagues.
 Do I need all the ‘problems’ or only those strictly relevant to coordination?
 This will not be demonstrated for a matter of space. See H&Z (2003) ‘Coordinated Bare Definites’ (p.459).
 See the ‘Activation Principle’ for further clarifications.
 Aren’t ‘vertebrato’ e ‘mammifero’ adjectives, and not nouns?
 I didn’t get the distinction made by Roodenburg between ‘strong’ and ‘weak’ CBNs.
 Actually, this is not necessarily the case, because a person can be considered both ‘a friend’ and ‘an enemy’, but that will be considered as an exception.