A descriptive study of brazilian offensive phrases

Ana C. Bastos-Gee*

*Universidade do Minho, Centro de Estudos Humanísticos, Braga, Portugal.

abastos@ilch.uminho.pt

 

ABSTRACT

In this paper I describe aspects of the grammar of DP-of-DP offensive phrases in Brazilian Portuguese, classifying the offensive words in three morphosemantic classes. The first part of the paper focuses on characteristics that unify the offensive words into one single class of expressive content constructions. I discuss properties such as: speaker orientation, impossibility of modification by a degree adverb, headedness of the phrases, reversibility, and strength of the determiners. The second part of the paper focuses on characteristics that support a classification of these offensive words in three classes: expressive abstract nouns, epithets, and swear words. I discuss properties such as: morphosemantic characteristics, gender agreement, and number agreement.

Keywords: Brazilian Portuguese; descriptive grammar; offensive phrases; expressive abstract nouns; epithets; swearwords.

 

RESUMO

Neste artigo descrevo aspectos da gramática de sintagmas ofensivos do tipo DP-de-DP em Português Brasileiro, classificando as palavras ofensivas em três classes morfossemânticas. A primeira parte deste artigo enfoca características que unificam as palavras ofensivas em uma classe única de construções com conteúdo expressivo. Discuto propriedades como: a orientação para o falante, a impossibilidade de modificação por um advérbio de grau, o núcleo dos sintagmas, a reversibilidade e a “força” dos determinantes. A segunda parte deste artigo enfoca características que apoiam uma classificação dessas palavras ofensivas em três classes: substantivos abstratos expressivos, epítetos e palavrões. Discuto propriedades como: características morfossemânticas, a concordância de gênero e a concordância de número.

Palavras chave: Portuguese Brasileiro; gramática descritiva; sintagmas ofensivos; substativos abstratos expressivos; epítetos e palavrões.

 

1. Introduction

This paper describes constructions with offensive nouns and nominalized adjectives, such as certain expressive abstract words with ironic interpretation, epithets and swearwords in Brazilian Portuguese.

(1) [A beleza/ galinha/ droga da Maria] fugiu.
The(fem) beauty/ hen/ drugof- the(fem) Mary vanished.
‘That lazy/ promiscuous/ piece of shit of Mary vanished’

Each one of the offensive nouns in (1) above is representative of a different class of offensive words in Brazilian Portuguese. The expressive abstract noun beleza ‘beauty’ is used ironically to describe Maria ‘Mary’ as ‘vain, snobbish or lazy’; the epithet galinha ‘hen’ is a pejorative way of characterizing a person who has many lovers; and the swear word droga ‘drug’ expresses speaker’s strong disapproval in a more general fashion, similarly to the English expression ‘piece of shit’, which I use to translate all members of this class.

The term expressive content is used here to refer to words and phrases that carry emotional content, such as anger, surprise, affection, etc. This is the case of the offensive words studied in this paper, which are used to express speaker’s anger, bother, or irony. In addition to the cases exemplified in (1), words with expressive content are often used as interjections or vocatives. I limit the scope of this paper to the cases in which they are used within a traditional nominal phrase that conforms to the schema DP₁-of-DP₂. In (2) I show the labels I use to refer to the different items in this construction.

(2) D₁ N₁ of-D₂ N₂
A droga da Maria
the drug of-the Mary
‘that piece of shit of Mary’

The term traditional nominal phrase is a neutral term used here to refer to a phrase that has a noun as its semantic head and to replace the term determiner phrase (DP), since recent studies of phrase structure have argued for the availability of additional phrases higher than DP in the nominal phrase (Ormazabal 1991, Ogawa 2001, Bastos-Gee 2011, among others). In (2) there is a traditional nominal phrase formed by two other traditional nominal phrases. The first noun in the linear order is the expressive word and its label is N₁. The determiner that immediately precedes N₁ is D₁. The second noun in the linear order is the semantic head of traditional nominal phrase, as I show in section 2.3, and its label is N₂. The second determiner in the structure is D₂ and it is often contracted with the preposition de ‘of’.

This paper has two goals. The first goal is to describe properties that are common to all constructions containing offensive words in Brazilian Portuguese. The second goal is to describe the properties that justify a classification of offensive words, based on morphosemantic properties and agreement patterns. The generative grammar framework, terminology and methodology are used to guide the description of the collected data.

The main language studied in this paper is Brazilian Portuguese; although, whenever available in the literature, I refer to comparable cases in other languages, especially Spanish. Brazilian Portuguese was chosen as object of study, because to the best of my knowledge there is no detailed description and classification of offensive words available in the literature of this language.^[1] The data was collected through grammaticality tests applied to speakers of Brazilian Portuguese, from different regions of Brazil, and also includes my own linguistic judgments.

This paper is organized in two parts. In section 2, I describe the properties that unify the offensive words into one large class. These properties are: speaker orientation, impossibility of modification by a degree adverb, headedness of the phrases, reversibility, and strength of the determiners. In section 3, I describe three different classes of constructions with offensive word, focusing on their morphosemantic characteristics, gender agreement, and number agreement.

2. Common properties of constructions with offensive words

As mentioned above, offensive words in Brazilian Portuguese can be classified in expressive abstract nouns, epithets and swear words. In this section, I discuss properties that are common to these three classes of offensive words, when they are within a DP-of-DP structure. These properties are speaker orientation, also referred to as main clause interpretation or widest scope, impossibility of degree modification, inverse headedness, interaction with the movement of other phrases out of the traditional nominal phrase, impossibility of reversibility, and the definiteness issue.

2.1. Offensive words are speaker oriented

Expressive abstract nouns, epithets and swear words are interpreted as a semantic contribution of the speaker of the sentence, but syntactically they are clearly within the limits of a traditional nominal phrase. This observation is not new, and it has been made for many languages. Just to cite a few cases, Aoun, Choueiri, and Hornstein (2001: 386) refer to this property of epithets in Lebanese Arabic as a “main clause” interpretation, and Aoun and Choueiri (2000:2–3) attest that some epithets in that language have an extra definite marker, a characteristic that distinguishes them from other nominal-internal items, but which reinforces the claim that they are items within a traditional nominal phrase themselves. Similarly, Potts (2005) observes for English that expressions carrying expressive content, such as expressive adjectives and epithets, have “widest scope”, and Huddleston and Pullum (2002:553) show that expressive adjectives in this language behave like strictly attributive adjectives, such as former and premier, in the sense that they must be prenominal and can co-occur with other adjectives in the traditional nominal phrase. Potts (2005) also shows that in German, expressive adjectives are case-marked just like all other adjectives.

This property of being at same time interpreted as part of the discourse-layer, on one hand, and internal to the traditional nominal phrase, on the other hand, is also present in Brazilian Portuguese.

(3) A Maria disse que o chefe mencionou que o beleza do motorista sumiu.
the Mary said that the boss mentioned that the beauty of-the driver vanished.
‘Mary said the boss mentioned that lazy driver vanished’

(4) A Maria disse que o chefe mencionou que o filho da mãe do motorista
the Mary said that the boss mentioned that the son of-the mother of-the driver
sumiu.
vanished.
‘Mary said the boss mentioned that son of a bitch of a driver vanished’

(5) A Maria disse que o chefe mencionou que o merda do motorista sumiu.
the Mary said that the boss mentioned that the shit of-the driver vanished.
‘Mary said the boss mentioned that piece of shit of a driver vanished’

On one hand, the negative opinion towards the driver in the sentences above is perceived as a contribution of the speaker, not as Mary’s or her boss’s. On the other hand, beleza ‘beauty’ (expressive abstract noun), filho da mãe ‘son of a bitch’ (epithet) and merda ‘shit’ (swearword) are clearly within the traditional nominal phrase. It is actually not true for Brazilian Portuguese that items with expressive content behave exactly like other modifying items within the traditional nominal phrase, since such constructions have some special properties, which will be discussed next. However, the fact that a determiner precedes them strongly suggests that the expressive nouns are within the limits of the traditional nominal phrase.

2.2. Degree modification is not permitted

Di Tullio and Saab (2006)^[2] classify constructions with epithets in Spanish in two classes: attributive and referential. In a nutshell, referential N₁ in Spanish appear within a definite verb-argument containing a full DP₂ or proper name as N₂ while attributive N₁ in Spanish appear within an indefinite predicative expression containing a bare noun as N₂. According to their description, it is possible to make a clear-cut distinction between these two types of constructions in Spanish, which means that some expressive words are intrinsically referential while others are intrinsically attributive. This is not the case for Brazilian Portuguese, as exemplified below with the swearword droga ‘drug’.

(6) Within verb-argument
Eu comprei adroga da casa/ uma droga de (uma) casa.
I bought the drug of-the house/ a drug of a house
‘I bought that piece of shit of a house.’

(7) Within predicative expression
(Uma/ que)droga de casa, essa do João!
A/how drug of house this of-the John
‘What a piece of shit of a house this one of John’s is!’

The very same offensive word that appears in (6) also appears in (7). The example in (6) complies to the above conditions for referential constructions in Spanish while the example in (7) complies with the above conditions for attributive constructions. Furthermore, the verb-argument containing the offensive word can be definite or indefinite, and N₂ can even be a bare noun in (6).

One of the properties that distinguishes between referential and attributive constructions in Spanish, discussed by Di Tullio and Saab (2006) is modification by degree adverbs. N₁ in referential constructions admits modification by degree adverbs in Spanish while N₁ in attributive constructions does not allow it. Some examples are the following:

(8) a. Referential: El muy tonto de Juan/ el muy burro de Juan
The very goofy of John/ the very donkey of John
‘John is such goofy’ ‘John is such idiot’
b. Attributive: *Una muy gran/ extraordinaria/ maravilla de mujer
a very big/ extraordinary/ wonder of woman
‘such big/ extraordinary woman’ and ‘such wonder of a woman’

In the example above, the words tonto ‘goofy’ and burro ‘donkey’ are modified by the degree adverb muy ‘very’ in a DP-of-DP structure.

The examples below show degree adverbs modifying adjectives in two different constructions in Brazilian Portuguese.

(9) a. Um homem muito/ tão inteligente conversou comigo hoje.
A man very/ so intelligent talked to-me today.
‘A very intelligent man talked to me today’.
b. Esse homem é muito/ tão inteligente.
This man is very/ so intelligent.
‘This man is very/ so intelligent’.

In Brazilian Portuguese, none of the three morphosemantic classes of offensive words allow modification by degree adverbs, as shown below.

(10) Expressive abstract noun
O (*muito/ tão) beleza do motorista sumiu.
the very/ so beauty of-the driver vanished.
‘the lazy driver vanished’

(11) Epithet
O (*muito/ tão) mosca morta do motorista sumiu.
the very/ so fly dead of-the driver vanished.
‘the stiff driver vanished’

(12) Swearword
O (*muito/ tão) merda do motorista sumiu.
the very/ so shit of-the driver vanished.
‘the piece of shit of a driver vanished’

The examples above are comparable to what Di Tullio and Saab (2006) call referential constructions, because the offensive word is within a definite verb-argument modifying a DP o motorista ‘the driver’, not a bare noun. Notice that even if we were to remove the first determiner of the constructions above, the result would still be unacceptable.

Modification by degree adverbs is not possible in indefinite predicative expressions modifying bare nouns either.^[3]

(13) Expressive abstract noun
Uma (*muito/ tão) beleza de motorista, esse mais velho.
a very/ so beauty of driver, this more old.
‘Such a very good driver, this old one.’

(14) Epithet
Uma (*muito/ tão) mosca morta de motorista, esse mais velho.
a very/ so fly dead of driver, this more old.
‘Such a very stiff driver, this old one.’

(15) Swearword
Uma (*muito/ tão) merda de motorista, esse mais velho.
a very/ so shit of driver, this more old.
‘Such a piece of shit of a driver, this old one’

Even if the indefinite D₁ were not present, these sentences would still be unacceptable with a degree adverb.

From this point on, I limit the scope of this paper to the argumental cases, both definite and indefinite, as exemplified in (6), since most of the restrictions on offensive content within predicative expressions may be a reflex of constraints on bare nouns in Brazilian Portuguese, and therefore only indirectly related to the expressive content. I leave this suggestion open to further research.

For the sake of completeness of this description, when it comes to intensifying the degree of the expressive content in Brazilian Portuguese, the only process available is the recursion of DPs with expressive content, which I refer to as a “re-cursing” process.

(16) Swearword/ swearword/ swearword
A merda da droga da bosta do motorista sumiu de novo.
The shit of-the drug of-the crap of-the driver vanished of again
‘The piece of shit of the driver vanished again.’

“Re-cursing” is possible for all three classes of expressive content, and mixing members of different classes is perfectly acceptable.

2.3. The second noun is the head

In DP-of-DP constructions, for instance, o filho do João ‘the son of John’s’ or o filho do mecânico ‘the son of the mechanic’, the traditional head of the structure is the first noun in the linear sequence, filho ‘son’. This is, however, not the case for DP-of-DP constructions with offensive words, in which the N₂ is the traditional head. ^[4] The classic tests to show the head of these constructions involve anaphor binding and subject-verb agreement. The results of these tests are presented below.

The following examples with anaphor binding show a contrast between sentences without expressive content in (17)-(18) and sentences with expressive content in (19).

(17) O João_j /o mecânico_i se_i/j machucou na porta.
The John/ the mechanic himself hurt in-the door
‘John/ the mechanic hurt himself in the door.’

(18) O filho_l do João_j/ do mecânico_i se_*i/*j/l machucou na porta.
The son of-the John/ of-the mechanic himself hurt in-the door
‘The son of John’s/ of the mechanic hurt himself in the door.’

(19) O filho-da-mãe do João_j/ do mecânico_i se_i/j/ machucou na porta.
The son-of-the-mother of-the John/ of-the mechanic himself hurt in-the door
‘The son of a bitch of John/ of the mechanic hurt himself in the door.’

Se ‘himself/herself’ is an anaphor. In (17), the anaphor can bind the c-commanding DP o João ‘John’ and o mecânico ‘the mechanic’. In (18), the anaphor can bind the c-commanding DP o filho do João/ do mecânico ‘the son of John/ the mechanic’ as a whole, but it cannot bind the second DP o João ‘John’/ o mecânico ‘the mechanic’. The second DP is more embedded within the traditional nominal domain, and it does not c-command the anaphor. The case in (18) is different from (19). In (19), the anaphor can bind the second DP o João ‘John’ and o mecânico ‘the mechanic’. One may conclude from these facts that the structure of (18) and (19) are different. In fact, these facts are compatible with a number of structural analyses proposed in the literature for other languages. Just to cite a few among them, there are analyses in which DP₁ and DP₂ are in a small clause configuration (Den Dikken 2006, Suñer 1990, and Español-Echevarría 1997), others in which both DP₁ and DP₂ are modifiers of an abstract head (Di Tullio and Saab 2006) and others in which DP₁ is a pre-nominal modifier of N₂ (Bastos-Gee 2011). It is beyond the scope of this paper to discuss each of them in detail.

The second test used to identify the head of DP-of-DP structures is subject-verb agreement. When the first and second DPs have different number values, the verb agrees in number with the head of the subject.

(20) Subject-verb agreement
O chefe desses taxistas estacionou/ *estacionaram na frente da
the boss of-these taxi-drivers parked(sing)/ parked(pl) in-the front of the
minha casa.
my house
‘the boss of taxi drivers parked in front of my house.’

In (20), the verb agrees in the singular with chefe ‘boss’.

In the case of DP-of-DP with offensive words, there are only a few cases in which the first and second DPs can have different number values. These restrictions will be described in section 3.3. For now, the cases that do allow the first and the second DPs to have different number values provide a piece of evidence for the second noun as the traditional head of the DP-of-DP subject.

(21) Subject-verb agreement
Essa merda desses taxistas *estacionou/ estacionaram na frente da
this shit of-these taxi-drivers parked(sing)/ parked(pl) in-the front of-the
minha casa.
my house
‘those piece of shit of taxi drivers parked in front of my house.’

In the above case, the verb must agree with N₂ taxistas “taxi drivers”, not with N₁ merda “shit”, which reinforces the conclusion that N₂ is the head of the traditional nominal phrase in (21).

2.4. The order between DP₁ and DP₂ is not reversible

Reversibility of the linear order is a property often found in DP-of-DP constructions. Moro (1997) and (2000) on his study of dynamic antisymmetry investigates phrases such as books of this type and this type of books. In Moro’s analysis, books and this type are generated in a small clause configuration, and either book or this type moves into the specifier of the preposition of.

(22) a. [… of [_SC books this type]
b. [books [of [_SC t this type]
c. [this type [of [_SC books t]

Suñer (1990) and Español-Echevarría (1997) argue that the possibility of inversion in nominal domains containing epithets in Spanish can best be analyzed by Moro’s style analysis. Some examples of inversion in Spanish phrases are the following:

(23) a. Un libro de porquería/ una porquería de libro
A book of garbage/ a garbage of book
‘that piece of shit of a book’
b. Un libro de mierda/ una mierda de libro
A book of shit/ a shit of book
‘that piece of shit of a book’

Di Tullio and Saab (2006) criticize the empirical coverage of their analysis. According to them, inversion is possible for attributive epithets only and there are exceptions even among members of the attributive class, as shown below.^[5]

(24) Una birria de cerveja/ *una cerveja de birria
a rubbish of beer/ a beer of rubbish
‘that piece of shit of a beer’

As for Brazilian Portuguese, inversion of the type described above is nearly inexistent and I could find only one case that allows it, as shown below.

(25) Um livro de merda/ uma merda de livro
A book of shit/ a shit of book
‘that piece of shit of a book’

Except for this one case, the order DP₁-of-DP₂ is not reversible at all in Brazilian Portuguese. This suggests that de merda ‘of shit’ may be a frozen expression in the language instead of the result of a syntactic process.

2.5. Determiners agree in “strength”

In this section, I describe a property of the determiners in DP-of-DP structure with offensive items: “strength agreement”. Before we go into this property, it is important to clarify that only high determiners can appear in constructions with expressive content. The following example shows high and low pre-nominal modifiers in BP.

(26) High determiners low determiners
As/ aquelas/ umas poucas/ muitas/ várias modelos viajaram.
The/ those/ a few/ many/ several models traveled
‘Few/ many/ several models traveled’.

From this point on, when I mention to the property of “strength” of determiners, I refer to high pre-nominal determiners only.

The term “strength” refers to the classification of determiners into strong and weak determiners. Strong determiners, such as demonstratives, induce definiteness effects while weak determiners, such as indefinite articles, do not induce definiteness effects. The classic test to show this distinction involves movement of wh-phrases out of the traditional nominal domain.

(27) Strong determiner: demonstratives
*De quem você rasgou [essa/aquela foto t_i ]?
Of whom you tore [this/that picture t_i]
‘Who is the person such that you tore apart this picture of him?’

(28) Weak determiners: indefinite, as well as low modifiers
De quem você rasgou [uma/ muitas/ várias foto(s) t_i]?
Of whom you tore [a/ many/ several picture t_i]
‘Who is the person such that you tore apart his picture?’

Strong determiners do not allow movement of wh-phrases out of the traditional nominal domain, as in (27) above, while weak determiners allow it, as in (28). As for the definite article, Brazilian Portuguese has two homophonous versions of it: a strong definite article and a weak definite article (see Bastos-Gee 2011 for discussion).

(29) Strong definite article
*De quem você rasgou [a foto preferida t_i]?
Of whom you tore the picture favorite t_i]
‘Who is the person such that you tore apart his favorite picture?’

(30) Weak definite article
De quem você rasgou [a foto t_i]?
Of whom you tore the picture t_i]
‘Who is the person such that you tore apart his picture?’

Speakers of Brazilian Portuguese see a contrast between (29) and (30). Under a semantic viewpoint, one of the differences between them is that in (29) we have a specific picture while in (30) the picture is less specific. The existence of two versions of the definite article within the same language is a common property among Romance languages (Torrego 1987, Ormazabal 1991, Vernaud and Zubizarreta 1992, Longobardi 1994, Ticio 2003, Tellier and Valois 1995, among others.) and the implications of it for the study of the definiteness effect has been extensively studied in the generative literature from both syntactic and semantic viewpoints.

The generalization for Brazilian Portuguese is that D₁ agrees with D₂ in “strength”. This generalization holds for argumental cases only, since all the predicative cases have a bare noun as N₂. To the best of my knowledge, the facts described in this section are new observations. Constructions with expressive nouns within the traditional nominal phrase show a definiteness effect when D₂ is a definite article or a demonstrative. This is true for all semantic types of expressive nouns.

(31) Eu despedi a/essa/*uma/*Ø gracinha da/ dessa modelo
I fired the/ this/ a/ Ø little-grace of-the/ of-this modelo
‘I fired that snobbish model’

(32) Eu despedi a/essa/*uma/*Ø mosca-morta da/ dessa modelo
I fired the /this/a/ Ø fly-dead of-the/ of-this modelo
‘I fired that stiff model’

(33) Eu despedi a/essa/*uma/*Ø merda da/ dessa modelo
I fired the /this/a/ Ø shit of-the/ of-this modelo
‘I fired that piece of shit of a model’

As illustrated in (31)-(33), if D₂ is a definite article or demonstrative, then D₁ must be a definite article or demonstrative. In this case, D₁ cannot be an indefinite article, and it cannot be Ø.

The results for a weak D₂ are shown below.

(34) Eu despedi a /*essa/ uma/*Ø gracinha de uma modelo
I fired the /this/a/ Ø little-grace of a modelo
‘I fired that snobbish model’

(35) Eu despedi a /*essa/ uma/*Ø mosca-morta de uma modelo
I fired the /this/a/ Ø fly-dead of a modelo
‘I fired that snobbish model’

(36) Eu despedi a /*essa/ uma/*Ø merda de uma modelo
I fired the /this/a/ Ø shit of a modelo
‘I fired that snobbisb model’

If D₂ is an indefinite article, then D₁ can be a weak definite article or an indefinite article. In this case, D₁ cannot be a demonstrative or Ø.

2.6. Summary of the section

In this section I discussed a number of properties that are commonly found in the literature of DP-of-DP constructions in other languages, especially Spanish. I showed that the three classes of offensive words in Brazilian Portuguese have many properties in common.

 

Table 1: Common properties

	Speaker-orientation	Degree adverbs	N2 as head	Reversibility	Strength “agreement”
Abstract nouns	Yes	No	Yes	No	Yes
Epithets	Yes	No	Yes	No	Yes
Swear words	Yes	No	Yes	No, with one exception	Yes

 

So far, this paper described properties that all offensive words share. In the next section, I provide a detailed description of each of the morphosemantic classes, pointing out exceptions, if needed, and emphasizing generalizations. I also show patterns of gender and number agreement.

3. Differences between classes of offensive words

There are three morphosemantic classes of offensive nouns in Brazilian Portuguese. The main characteristics that distinguish them are whether they were originally abstract nouns, concrete nouns or interjections. As discussed previously, whether the expressive word modifies an individual or a group has been considered one the most important properties to the characterization of expressive content in Spanish. Different from Spanish, the ability to combine with an individual or a group does not contribute to establish a clear-cut distinction between two classes in Brazilian Portuguese. There is rather a gradation, which I describe as main tendencies for each class in addition to pointing out exceptions, if needed.

3.1. Morphosemantic properties

Brazilian Portuguese has quite a few nouns that can carry expressive content in the construction studied here, as well as some nominalized adjectives. Before I describe the syntactic and morphosyntactic properties that justify a three-way classification, I discuss the morphosemantic basis to this classification of offensive words into expressive abstract nouns, epithets and swear words. Let us start with expressive abstract nouns.

(37) A minha irmã/ a Maria/ a minha gata/ a minha chave é uma beleza/ gracinha.
The my sister/ the Mary/ the my cat/ the my key is a beauty/ little-grace
‘my sister/ Mary/ my cat/ my key is very good.’

(38) Expressive abstract nouns
A beleza/ gracinha/ lindeza da minha irmã/ Maria/ da minha gata/ da minha chave
the beauty/ little-grace/ beauty of-the my sister/ Mary/ of-the my cat/ of-the my key
desapareceu de novo
vanished again.
“that no-good of a sister of mine/Mary/ a cat of mine/ a key of mine vanished again”

Abstract nouns with expressive content, such as in (37) and (38), are usually combined with very specific classes of nouns. The ones presented above can characterize persons, animals and objects. In (37), they are not in a prenominal position within the traditional nominal phrase, and their meaning is positive. When in a definite traditional nominal phrase as in (38), these abstract nouns can only convey irony; they express the property of being ‘no good’. When combined with human individuals, the property of being ‘no good’ is sometimes interpreted as ‘vain, snobbish or lazy’, depending on the context.

One special note should be made with respect to the presence of diminutive or augmentative morphemes in these constructions. Diminutives and augmentative morphemes are often used in Portuguese to convey expressive content, and they are often added to abstract nouns (and epithets, as well). For instance, the abstract word beleza ‘beauty’ is used with feminine diminutive for women, a belezinha ‘the little pretty’ and with masculine augmentative for men, o belezão ‘the big pretty’.

Additional examples of abstract nouns that can only characterize very specific types of nouns are asco ‘disgust’ and fiasco ‘failure’, which combine with situations and events, and with food, respectively.

(39) a. O fiasco da festa/ apresentação/ campanha me deu dor de cabeça.
the failure of-the party/ presentation/ campaign me gave acke of head.
“the failure of a party/ presentation/ campaign gave me a headacke.”
b. O asco da sopa/ comida me deu dor de cabeça
the disgust of-the soup/ food me gave acke of head.
“the disgust of a soup/ food gave me a headacke.”

A few expressive abstract nouns, like amorzinho ‘little love’ and maravilha ‘wonder’, can only be combined with bare nouns in the position of N₂. In this case, they have a kind of ironic interpretation when used in a definite/ demonstrative traditional nominal phrase, as shown below. This ironic interpretation is better translated by the phrase “the so called” in English.

(40) a. O amorzinho de menina mentiu para mim.
the little-love of girl lied to me.
“that so-called love of a girl lied to me.”
b. Essa maravilha de cidade tem altos índices de criminalidade
this wonder of city has high rates of criminality
“this so-called wonder of a city has high rates of criminality.”

As mentioned above, this paper focus on expressive nouns, but it is important to notice that a large number of attributive adjectives can be “nominalized” and occupy the position of N₁ in the structure. The following are just a few examples with attributes that are originally negative and positive.

(41) Nominalized adjectives
A horrorosa/ antipática da minha irmã/ Maria sumiu de novo.
the ugly / unpleasant of-the my sister/ Mary disappeared again.
“that lazy/ antisocial of a sister of mine/ Mary disappeared again.”

(42) Nominalized adjectives
A bonita/ inteligente/ simpática da minha irmã/ Maria sumiu de novo.
the pretty/ intelligent/ pleasant of-the my sister/ Mary disappeared again.
“that lazy/ dumb/ antisocial of a sister of mine/ Mary disappeared again.”

If the nominalized adjective denotes a negative attribute, its original meaning remains the same, but if it denotes a positive attribute, then the ironic interpretation is the only one possible.^[6]

The most salient property of the members of this class is the ironic interpretation that is present when they are in a traditional nominal phrase that is the argument of a verb, especially definite arguments. Expressive abstract nouns lose their ironic interpretation within a predicative expression, as well as in post-nominal position, as will be shown below.

Opposing to the class of abstract nouns, the class of epithets has nouns that were originally concrete nouns, as exemplified below.

(43) Epithets
A mosca-morta/ galinha/ banana/ laranja da minha irmã/ Maria parou de trabalhar.
The fly-dead/ hen/ banana/ orange of-the my sister/ Mary stopped of work
“That stiff/ promiscuous/ wimp/ fool of a sister of mine/ Mary stopped working.”

(44) Epithets
*A mosca-morta/ galinha/ banana/ laranja minha impressora parou de trabalhar.
The fly-dead/ hen/ banana/ orange my printer stopped of work
“That stiff/ promiscuous/ wimp/ fool of a printer of mine stopped working.”

Epithets, in the strict sense I use in this paper, were originally concrete nouns used metaphorically to persons to describe an attributed quality. The contrast between (43) and (44) shows that those epithets and a large number of others collected during my research can only and exclusively be combined with human individuals, with exceptions for pets and pet-objects when they are given anthropomorphic characteristics by the speaker, and only if they can be characterized by the property described by the epithet. One curious aspect of epithets is that each of them has a very specific meaning that sometimes can no longer be associated with the literal meaning of the term that they have originated from. In the examples above, for instance, mosca-morta ‘dead fly’ is someone who is stiff and not lively, galinha ‘hen’ is a promiscuous person, banana ‘banana’ is a wimp, and laranja ‘orange’ is a fool who was blamed for something he or she did not do, especially in a money laundering operation.

Additional members to the list of epithets, in strict sense, are the following ones, accompanied by literal translations only: animal ‘animal’, anta ‘tapir’, babaca ‘idiot’, babão ‘fool’, bacana ‘rich, cool’, barata tonta ‘flustered cockroach’, bundão ‘big bottom’, burro(a) ‘donkey’, cachorro(a) ‘dog’, canalha ‘scoundrel’, chato(a) ‘pubic louse’, corno(a) ‘cheated man’, doido(a) ‘crazy’, escroto(a) ‘scrotum’, filho(a)-da-mãe ‘son of the mother’, filho(a)-da-puta ‘son of the whore’, idiota ‘idiot’, mula ‘donkey’, pamonha ‘cornbread’, pentelho(a) ‘pubic hair’, peste ‘pest’, porre ‘drunk’, and puto(a) ‘whore’. There is also the epithet gato(a) ‘cat’, which does not convey an offensive characterization per se, but it is slightly vulgar, meaning ‘sexually attractive’. The epithets above can be used generally for any person, but there are some epithets that are specially used to express prejudiced views against minorities based on gender, ethnicity and/or sexual orientation. All of them are epithets in the strict sense used here. The ones that are offensive to homosexuals are only combined with male persons, such as boiola ‘gay’, cuzão ‘big bottom’, frutinha ‘little fruit’, and veado ‘deer’; the ones that are offensive to women are only combined with female persons and usually originate from animal designations, perua ‘female turkey’, piranha ‘piranha’, vaca ‘cow’, cadela ‘bitch’, égua ‘mare’, etc.; and the ones that are offensive to people who live in the countryside are matuto(a) ‘countryside person’, caipira ‘countryside person’, zé bedeu ‘nosy Jack’, zé ruela ‘smallroad Jack’, bicho-do-mato ‘beast of the jungle’, etc.

It is not my intention to provide a complete glossary of epithets in Brazilian Portuguese. My goal is to show that this class is large and specialized. The most salient property of this class is that these offensive words were originally concrete nouns, which were used in a figurative or metaphoric sense to convey expressive content towards human individuals. The great majority of epithets cannot be combined with bare nouns, which in Brazilian Portuguese are used to refer to kinds of things. Nowadays, they have an expressive meaning, which is not always transparent by looking into their literal definition.

Let us now move on to the swearwords.

(45) Swear words
A merda/ bosta/ porcaria/ porra/ droga da minha irmã/ Maria parou de trabalhar.
the shit/ crap/ garbage/ semen/drug of-the my sister/ Mary stopped of work
All: “the piece of shit of a sister of mine/ Mary stopped working.”

(46) Swear words
A merda/ bosta/ porcaria/ porra/ droga da minha impressora parou de trabalhar.
the shit/ crap/ garbage/ semen/drug of-the my printer stopped of work
All: “the piece of shit of a printer of mine stopped working.”

Finally, swear words shown in (45) and (46) were originally interjections of anger. The ones shown above were originally concrete nouns, but being concrete or abstract is not the most important property, since they are now somewhat empty semantically. In a DP-of-DP structure, they can be freely used for any type of noun – person, animal, object, situation, etc. – and they all mean pretty much the same thing, i.e. they are all used to express speakers strong negative feelings toward something or someone, similarly to the English ‘piece of shit’ expression that I used to translate them.

The following table summarizes the semantic relation between type of offensive noun and type of N_2.

 

Table 2: N1 versus N2

N2 N1	Human individual	Anthropo-morphized pets/ pet-objects	animals	objects	places	concepts	Kinds (bare nouns)
Abstract nouns	Some	Some	Some	Some	Some	Some	Some
Epithets	Yes	Yes	No	No	No	No	A few
Swear words	Yes	Yes	Yes	Yes	Yes	Yes	Yes

 

As mentioned previously, abstract nouns can only be combined with N₂ that have specific semantic properties, for instance, either human individuals or situations. Epithets can only be combined with human individuals. Swear words can be combined with any kind of noun, including people, animals, objects, concepts, etc. There are no semantic restrictions on what swear words can combine with.

A syntactic piece of evidence for the distribution presented above can be seen below with respect to post-nominal occurrences of expressive content.^[7]

(47) Abstract nouns and epithets
Eu tenho um advogado beleza/ gracinha/ mosca-morta/ galinha/
I have a lawyer beauty/ little-grace/ dead-fly/ hen/
banana/ laranja
banana/ orange
“I have a lawyer, who is nice/ nice/ stiff/ promiscuous/ wimp/ fool”

(48) Swear words
*Eu tenho um advogado merda/ bosta/ porcaria/ porra/ droga
I have a lawyer shit/ crap/ garbage/ semen/ drug
All: “I have a piece of shit of a lawyer”

Post-nominal occurrences of expressive words are different from reversibility, because there is no DP-of-DP structure in the examples above. There is only one nominal domain and the expressive nouns are in an adjective-like position. Swearwords, due to their origin as interjections, cannot occupy adjective-like positions within the traditional nominal phrase. The examples in (47)-(48) show a clear separation between expressive abstract nouns/ epithets, on one hand, and swearwords, on the other hand.

To summarize to information in this section, one can say that generally speaking, swearwords can combine with any kind of N₂, epithets can only combine with human individuals or animals/ pet-objects with anthropomorphic properties, and expressive abstract words are a heterogeneous class, in which some expressive nouns can only combine individuals, others can only combine with objects, and so on.

3.2. Gender agreement mismatch

With respect to agreement between DP₁ and DP₂, the paradigm depends on the morphosemantic class of the expressive word and on morphosyntactic properties of N₂. In general lines, agreement between DP₁ and DP₂ is optional for swear words, but obligatory for expressive nouns and epithets. There are a few exceptions among epithets, which I describe case by case.

Before I discuss the actual data, I will review the basic properties of gender agreement in the Brazilian traditional nominal phrase. As in other Romance languages, determiners, nouns and adjectives agree in gender within the traditional nominal phrase in Brazilian Portuguese, as exemplified below.

(49) O pat-o pret-o versus a pat-a pret-a
The(masc) duck-masc black-masc the(fem) duck-fem black-fem
‘The black duck’

In the above example, the noun gato(a) ‘cat’ and the adjective preto(a) ‘black’ are both biform, i.e. they have two distinct inflectional forms for gender. In the above example, the distinction is between the inflectional morpheme –o, masculine, and the inflectional morpheme –a, feminine. Brazilian Portuguese, like other Romance languages, distinguishes between masculine and feminine only, and when the reference to mix-gender groups is necessary, the suffix –o is used.^[8] The examples in (50)-(54) show different classes of nouns and ways of identifying gender in Brazilian Portuguese.

(50) heteronyms: o homem ‘the(masc) man’ and a mulher ‘the(fem) woman’

(51) inflectional biforms: o pato ‘the(masc) duck’ and a pata ‘the(fem) duck-fem’

(52) uniforms: o/a pianista ‘the(masc)/ (fem) pianist’

(53) epicene: a zebra macho/ fêmea ‘the(fem) zebra male/ female’

(54) grammatical gender only: a mesa ‘the(fem) table’

One of the most common ways of distinguishing natural gender in Brazilian Portuguese is through the semantic relation between two nouns with different stems, as in (45). The cases in which natural gender is expressed by inflectional forms are exemplified in (46) and referred to as biform nouns. Different from those, Brazilian Portuguese has some nouns that have only one form, but natural gender can be identified by the masculine or feminine determiners that accompany them, as in (52). The traditional grammar also mentions epicene nouns like the one in (53), which are usually animal common nouns with only one form for masculine and feminine, but to which the terms macho/fêmea ‘male/female’ can be added in order to disambiguate the biological gender of the animal. The last two cases are considered contextual disambiguation of the natural gender, not morphological properties. Importantly, the gender of the epicene noun itself does not change after the terms macho/fêmea ‘male/female’ are added to the phrase. Finally, Brazilian Portuguese does not have neuter gender; even nouns that do not express biological gender are classified as masculine or feminine, as in (54) above.

Also relevant to my description is that some adjectives in Brazilian Portuguese are biform for gender and some are uniform, as exemplified below.

(55) Homem bonit-o/ útil versus mulher bonit-a/ útil
man pretty-masc/ useful woman pretty-fem/ useful
‘handsome/ useful man’ ‘pretty/ useful woman’

In (55) the adjective bonito(a) ‘pretty’ if biform, i.e. it has two forms for masculine and feminine while the adjective útil ‘useful’ has only one.

The properties described above for gender in nouns and adjectives are important, because nouns with expressive content are a mixed class of lexical items. They have some properties of nouns and some properties of adjectives. I now describe the main data regarding gender agreement in constructions with expressive content.

One of the most fascinating pieces of data regarding gender agreement in constructions with expressive content is the so-called “gender mismatch”, which consists of apparent gender agreement between D₁ with N₂, which ignores N₁ although it is syntactically closer to D₁. This phenomenon has been observed for French by Milner (1978) and for Spanish by Di Tullio and Saab (2006), Di Tullio and Suñer (2004), among others. In Spanish, differently from Brazilian Portuguese, the gender mismatch is only possible for what they call referential epithets. This restriction is not found in Brazilian Portuguese, where the apparent gender mismatch is possible for nouns belonging to any of the three classes of expressive nouns found in this language, as shown below.^[9]

(56) Abstract noun/ epithet/ swear word
O beleza/ banana/ merda do meu irmão fugiu.
the(masc) beauty(fem)/banana(fem)/shit(fem) of-the(masc) my brother vanished.
‘That promiscuous/ wimp/ fool of a brother of mine vanished.’

In Brazilian Portuguese, this apparent gender mismatch targets a very specific group of nouns: they are all expressive nouns, feminine in their literal meaning, and ending with the suffix –a (or –e, such as peste ‘pest’ and gilete ‘gillette’). The morphological endings –a, –e,
–ista and –nte are common endings for words that belong the group of nouns that are uniform for gender, as exemplified above in (52). These are nouns, in which the natural gender is identified by the determiner that accompanies them. It is important to notice that the masculine gender of D₁ above is still dependent of the gender of N₂.

There are two exceptions to the morphological group described above: anta ‘tapir’ and mula ‘donkey’. These two epithets are always feminine and agree in feminine gender with their determiners.

There are no cases of gender mismatch between N₁ and D₁ with masculine N₁ as exemplified below. ^[10]

(57) Swear word/ epithet/ epithet
O/ *A caralho/ animal/ asno d-a Maria
the(masc)/the(fem) dick(masc)/animal(masc)/donkey(masc) of-the(fem) Mary
sumiu.
vanished.
‘That dick/ animal/ idiot of Mary vanished.’

The example in (57) shows that if N₁ is masculine, a feminine D₁ cannot precede it. This means that the so-called gender mismatch only targets feminine nouns. It is, however, a fact that the great majority of expressive nouns in Brazilian Portuguese are feminine.

Among the epithets, there are many that are biform nouns and many that were originally biform adjectives used now expressively. The epithet burro ‘donkey’ is an example of these cases.

(58) Epithet
a. O burr-o d-o João sumiu.
the(masc) donkey-masc of-the(masc) John vanished.
‘That stupid John vanished.’
b. *a burr-a d-o João sumiu.
the(fem) donkey-fem of-the(masc) John vanished.
‘That stupid John vanished.’

(59) Epithet
a. *O burr-o d-a Maria sumiu.
t he(masc) donkey-masc of-the(fem) Mary vanished.
‘ That stupid Mary vanished.’
b. a burr-a d-a Maria sumiu.
the(fem) donkey-fem of-the(fem) Mary vanished.
‘That stupid Mary vanished.’

Biform epithets do not show gender mismatch, but the gender of N₁, as well as D₁, is clearly dependent on the gender of N₂.

One exception to rule that all epithets have obligatory agreement between DP₁ and DP₂ is a small group of noun-adjective compound epithets. These epithets have optional agreement. To the best of my knowledge this is a new observation.

(60) A mosca-morta/ barata-tonta do João sumiu.
the(fem) ly-dead/ flustered cockroach of-the(masc) John vanished.
‘That stiff/ flustered John vanished.’

(61) O mosca-morta/ barata-tonta do João s umiu.
the(masc) fly-dead/ flustered cockroach of-the(masc) John vanished.
‘That stiff/ flustered John vanished.’

This small class of epithets behaves similarly to swearwords as for the gender agreement, in the sense that they show optional agreement, not mandatory.

The class of swear words behaves differently from epithets and abstracts nouns. To the best of my knowledge, this observation is also new, recently made on earlier versions of this work (Bastos-Gee 2011). Consider (62)-­(64).

(62) Swear word
O merda d-o homem/ pat-o sumiu.
the(masc) shit(fem) of-the(masc) man/ duck-masc vanished
‘that piece of shit of a man/ duck vanished’

(63) Swear word
A merda d-o homem/ pat-o sumiu.
the(fem) shit(fem) of-the(masc) man/ duck-masc vanished
‘that piece of shit of a man/ duck vanished’

(64) Swear word
*O merda d-o carro sumiu.
the(masc) shit(fem) of-the(masc) car(masc) vanished
“that piece of shit of car vanished”

(65) Swear word
A merda d-o carro sumiu.
the(fem) shit(fem) of-the(masc) car(masc) vanished
“that piece of shit of car vanished”

In the above examples, gender agreement between D₁ and N₂ (i.e. the so-called “gender mismatch” between D₁ and N₁) is only possible if N₂ has natural gender, but not possible if N₂ has only grammatical gender. D₁ can be masculine in (62) agreeing with the nouns homem ‘man’ and pato ‘duck’, but it cannot be masculine in (64) agreeing with carro ‘car’. Notice that this issue does not arise with epithets, because they can only be combined with human individuals, i.e. elements with natural gender, to start with.

It is important to clarify that this is not a case of “semantic” agreement in the traditional sense of “gender sylepsis” between D₁ and N₂; this is a real syntactic issue. One way of testing for the issue is to use an epicene, i.e. a noun that is either masculine or feminine, but allows natural gender disambiguation through the adjectives macho ‘male’ and fêmea ‘female’. One example is provided below.

(66) A cobra macho sumiu.
the(fem) snake male vanished
“the male snake vanished.”

The word cobra ‘snake’ is grammatically feminine, and the adjective macho ‘male’ is used to contextually identify natural gender when the distinction is relevant, for instance, to contrast it with a female snake in the same context. The test has the following format. In traditional nominal phrases with expressive content, if D₁ agrees with the grammatical gender of N₂, it should be feminine; if D₁ agrees with the contextually salient masculine gender, it should be masculine.

(67) Swear word
A/ *o merda d-a cobra macho sumiu.
the(fem)/ the(masc) shit the(fem) snake male vanished
“that piece of shit of a male snake vanished.”

As shown above, D₁ cannot agree with the semantic idea of masculine; the only agreement possible is the syntactic agreement. This shows that the gender mismatch is a syntactic phenomenon, and not a semantic, contextual phenomenon.

Another way in which swearwords are different from epithets and abstract nouns is that the so-called “gender mismatch” is optional for swear words, even when N₂ has natural gender, as shown in (62)-(65) above. As can be seen in the examples with swear words, there is always the possibility of D₁ and N₁ agreeing internally within their own traditional nominal phrase, and having a different gender from N₂, even when N₂ is human. There is no such optionality with epithets and abstract nouns, where either D₁ agrees with N₂ in gender or both D₁ and N₁ agree with N₂ in gender when N₂ has natural gender, exception made to noun-adjective compounds, as mentioned above.

The following table summarizes my proposal for the gender feature of nouns with expressive content in Brazilian Portuguese.

 

Table 3: gender agreement

Type of expressive noun	D1-N1 gender ‘mismatch’	DP1-N2 gender agreement
Feminine abstract nouns, ending –a and –e beleza ‘beauty’	Yes	Mandatory, if N2 has natural gender
Feminine epithets, ending –a and –e banana ‘banana’	Yes	Mandatory
Exceptions: Anta ‘tapir’, mula ‘donkey’	No	No
Type of expressive noun	D1-N1 gender ‘mismatch’	DP1-N2 gender agreement
Masculine abstract nouns, epithets, and swear words Fiasco ‘failure’, asno ‘donkey’, and caralho ‘dick’	No	No
Biform epithets burro(a) ‘donkey/ idiot’	No	Mandatory
Noun-Adjective compound epithets mosca-morta ‘dead fly’	Optional	Optional
Swear words merda ‘shit’	Optional, if N2 has natural gender No, if N2 has grammatical gender	Optional, if N2 has natural gender No, if N2 has grammatical gender

 

In the next section, I describe the properties of gender agreement.

3.3. Number agreement mismatch

Before I discuss the actual data, I will review the basic properties of number agreement in the Brazilian traditional nominal phrase. In Brazilian Portuguese, like other Romance languages, determiners, adjectives and nouns agree in number within the traditional nominal phrase.

O pato preto versus o-s pato-s preto-s
The(sing) duck(sing) black(sing) the-pl duck-pl black-pl
‘The black duck(s)’

In Brazilian Portuguese, plural nouns, adjectives and determiners are marked with the inflectional suffix –s (and its allomorphs) that is added to the stem of the word, after the feminine suffix if there is one. Singular forms are indicated by the absence of the morpheme –s.^[11]

When it comes to number agreement in traditional nominal phrases with expressive content, the paradigm is considerably simpler than the one presented for gender agreement: Number agreement is obligatory when N₁ is an abstract word or an epithet, and optional when N₁ is a swear word. The following examples illustrate definite traditional nominal phrases but the generalization is also true for indefinite traditional nominal phrases in argumental positions. Also, the distinction between uniform/ biform epithets does not play a role here.

(69) Abstract noun
A-s beleza-s/ *a beleza d-a-s modelo-s
The(fem)-pl beauty-pl/ the(fem) beauty of-the(fem)- pl model-s
viajaram.
traveled.
‘those snobbish models traveled.’

(70) Epithet
A-s banana-s/ *a banana d-a-s modelo-s
The(fem)-pl banana-pl/ the(fem) banana of-the(fem)- pl model-s
viajaram.
traveled
‘those wimpy models traveled.’

(71) Swear word
A-s merda-s/ a merda d-a-s modelo-s
The(fem)-pl shit-pl/ the(fem) shit of-the(fem)- pl model-s
viajaram.
traveled
‘those shitty models traveled.’

In (69)-(70) above, abstract nouns, epithets and their respective determiners must agree with N₂ in number. In (71), the swear words display optional number agreement in the same way they display optional gender agreement, as discussed in the previous section. One empirical difference between the gender and the number of constructions with swear words is that there is no “number mismatch”, in the sense I used the term “mismatch” for gender.

A piece of evidence for it is that there are subtle differences in the interpretation of the traditional nominal phrases containing singular swear words and plural swear words in (71) above. In a merda das modelos ‘the shit of the models’, the speaker is upset with the models as a group. In as merdas das modelos ‘the shits of the models’, the speaker attributes the negative property to each of the members of the group. This leads me to the conclusion that in both cases, D₁ agrees internally with N₁ and this agreement is semantically relevant for interpretation.

The following table summarizes my proposal for the number feature of nouns with expressive content in Brazilian Portuguese.

 

Table 4: number agreement

Type of expressive noun	D1-N1 number ‘mismatch’	DP1-N2 number agreement
Expressive abstract nouns beleza ‘beauty’,	No	Yes, mandatory
Epithets, including Noun-Adjective compounds banana ‘banana’, mosca-morta ‘dead-fly’	No	Yes, mandatory
Swear words merda ‘shit’	No	Optional

 

The partial conclusion for this section is that number agreement in traditional nominal phrases with expressive content depends on the semantic type of the expressive N₁. Expressive abstract nouns and epithets show mandatory agreement with N₂, while swear words show optional agreement with N₂, since their number value affects interpretation.

4. Final remarks

In this paper, I described a number of the properties of DP-of-DP constructions containing nouns that carry expressive content, in order to establish differences between this type of DP-of-DP construction and other kinds of DP-of-DP constructions. Additionally, I briefly compared the properties found in Brazilian Portuguese to the properties found in the literature about epithets and other offensive words in other languages, especially Spanish. This paper also includes a detailed description of three classes of offensive words in Brazilian Portuguese, as for their morphosemantic properties and the properties of the noun they modify. This classification of offensive words into three classes was crucial to describe complex patterns of gender and number agreement within the traditional nominal domain.

 

5. References

Aoun, J. and L. Choueiri. (2000). Epithets. Natural Language and Linguistic Theory 18, 1-39.         [ Links ]

Aoun, J., L. Choueiri and N. Hornstein. (2001). Resumption, Movement and Derivational Economy. Linguistic Inquiry 32, 371-403.         [ Links ]

Bastos-Gee, Ana C. (2011). Information structure within the traditional nominal phrase: the case of Brazilian Portuguese. University of Connecticut: Doctoral Dissertation.         [ Links ]

Câmara Jr., Joaquim Mattoso. (1984). Estrutura da língua portuguesa. 14. ed. Petrópolis, Vozes.         [ Links ]

Den Dikken, Marcel. (2006). Relators and linkers: The syntax of predication, Predicate Inversion, and copulas. Cambridge, MA: MIT Press.         [ Links ]

Di Tullio, A. and Suñer, A. (2004). Los ‘nombres de cualidad’ en la estructura del SD. Presented at the XXIV Congrès International de Linguistique et de Philologie Romanes (CILPR). Aberystwyph, August 1-6.         [ Links ]

Di Tullio, Ángela and Andrés Saab. (2006). Dos clases de epítetos en el español: sus propiedades referenciales y distribución sintáctica. Proceedings of the XIV Congreso Internacional de la Asociación de Lingüística y Filología de América Latina (ALFAL), Universidad Autónoma de Nueva León.         [ Links ]

Español-Echevarría, M. (1997). Two aspects of the sentential syntax of N/A of a N DP’s: predicate raising and subject licensing. In: A. Schwegler, B. Tranel and M Uribe-Exteberría (eds.) Romance Linguistics: Theoretical perspectives. Amsterdam: John Benjamins.         [ Links ]

Huddleston, Rodney and Geoffrey Pullum. (2002). The Cambridge Grammar of the English Language. Cambridge University Press.         [ Links ]

Kehdi, Valter. (1990). Morfemas do português. São Paulo: Ática, Série Princípios.         [ Links ]

Macambira, José Rebolças. (1992). Português Estrutural. São Paulo: Pioneira.         [ Links ]

Milner, Jean-Claude. (1978). De la syntaxe à l’interprétation: Quantités, insultes, exclamations. Paris: Éditions du Seuil.         [ Links ]

Moro, Andrea. (1997). The Raising of Predicates: Predicative Noun Phrases and the Theory of Clause Structure. Cambridge: Cambridge University Press.         [ Links ]

Moro, Andrea. (2000). Dynamic Antisymmetry, MIT Press, Linguistic Inquiry Monograph 38.         [ Links ]

Ogawa, Yoshiki. (2001). A Unified Theory of Verbal and Nominal Projections. Oxford: Oxford University Press.         [ Links ]

Ormazabal, Javier. (1991). Asymmetries on wh-movement and some theoretical consequences. Ms., Storrs: University of Connecticut.         [ Links ]

Potts, Chris. (2005). Expressive content as conventional implicature. Proceedings of the Thirty-Third Annual Meeting of the North East Lingusitic Society. BookSurge Publishing, December 7.         [ Links ]

Saab, Andrés. (2006). Concordancia ad sensum y elipsis nominal en español: Un análisis morfosintáctico. Revista de Lingüística y Literatura, Universidad Nacional del Comahue, n. 34, p. 45-63, 2006.         [ Links ]

Selkirk, E. (1977). Some remarks on noun phrase structure. In: P. Culiver, T. Wasow and Akmajian (eds.), Formal Syntax. New York: Academic Press, 285-316.         [ Links ]

Suñer, A. (1990). La predicación secundaria en español. Doctoral dissertation. Universidad Autónoma de Barcelona.         [ Links ]

 

Notes

^[1] This paper results from the development of many descriptive aspects originally reported in Bastos-Gee 2011, and presents new facts and new generalizations.

^[2] I refer the reader to Saab (2006) and Di Tullio and Suñer (2004) for developments in the analysis and description of the data in Spanish.

^[3] Abstract words loose their ironic interpretation in predicative expressions, although their content is still expressive. Also, there are very few epithets that can take a bare noun as their N₂. The epithet mosca morta ‘dead fly’ is an exception.

^[4] Among the first studies on reverse headedness is Selkirk (1977), which studies a type of pseudopartitive construction in sentences such as “She broke/ drank a bottle of wine”. The verbs to break and to drink show different semantic restrictions when selecting their complements, the verb to break selects a solid complement and the verb to drink selects a liquid complement. In Selkirk’s (1977) analysis, bottle is the head of the complement of to break in the “container reading” but wine is the head of the complement of to drink in the “content reading”. To account to the difference between container and content readings, she proposes that each reading will correspond to a different structure.

^[5] All translations from Spanish are my own.

^[6] A large number of attributive words are listed as both adjective and noun in glossaries and dictionaries. Therefore, even if I intend to have a list of nouns only, it is inevitable to have some nominalized adjectives among them; some are expressive abstract nouns and others, epithets. The most reliable criterion to separate expressive abstract nouns from epithets is whether they are originally concrete or abstract.

^[7] Notice that in a post-nominal position, expressive abstract nouns lose the “bothering” interpretation and can be interpreted as a positive aspect.

^[8] This analysis is actually controversial in traditional and structuralist grammars. There are two lines of analysis: Câmara Jr. (1984), Macambira (1987, 1992), among others, argue that the morpheme –a is an inflectional morpheme for feminine and –o is a neutral form that indicates the class of the noun (thematic vowel). Kehdi (1990), among others, argues that there is a true opposition between the morpheme –a, feminine, and the morpheme –o, masculine.

^[9] I show later that whether N₂ is human plays a role in this paradigm, because [+human] individuals have natural gender.

^[10] Thanks to Alberto Guerreiro (p.c.) who pointed out to me the colloquial neologism a caralha ‘the(fem) dick(fem)’, which derives a feminine form of the masculine swear word. His suggestion is that nouns that do not conform to the general rule for biform nouns might become regular in the historical evolution of the language.

^[11] In non-standard Brazilian Portuguese, the plural marker is usually retained in the first word in the traditional nominal phrase in the linear order, which is usually the determiner, but not exclusively.