Few subjects have been as difficult 10 explore dispassionately in contemporary Cuba as the subject of racial or ethnic ditlerences. On the one hand the revolutionary governmem of Fidel Caslro insists officially that ethnic differences no longer play an important role in public or private decisionmaking. The official position is that race is irrelevant - or at least of minor significance - in the construction of the socialist state, After 1959 the revolution sought to establish a state that deliberately privileged the lower orders of society, most of whom were black. In the rhetoric of the revolutionary government all forms of discrimination are legally prohibited (Lopez Valdes, 1973:6-14). Indeed, when the revolutionary Cubans finally wrote a new constitution in 1975 they made discrimination a violation of one's constitutional rights. Article 41 of the new Constitution of the Republic of Cuha declares that ‘discrimination because of race, calor, sex, or national origin is forbidden and is punished by law’ (1976: 13). This section mandated equal rights for all citizens. Formal institutional barriers to equality for any group were therefore removed by law. In practice, however, this constitutional equality does not exactly work itself out dearly or consistently. But matters of public policy and private action have always been highly contested and constitute an extremely ambiguous sphere of operations. Within the private domain the state is powerless 10 change the deeply held attitudes of some of its citizens, and these attitudes are especially important for matters of race and ethnicity. The degree to which citizens uphold the law is based on perceptions of self interest. When conformity brings greater reward than nonconformity, then citizens conform. At the same lime, intellectual traditions may not reflect daily conduct at the levels of public and private spheres (Helg, 1990).

While the government cannot regulate private racial and ethnic considerations in daily life, it subordinates these considerations 10 those of class and social justice. By promoting the goal of an egalitarian society, the government projected an ideal that its supporters were obliged 10 accept publicly. Elizabeth Sutherland observed in the 1960s after spending a summer traveling and working with the Cubans that ‘racism in the sense of the subordination of one racial group by another for the benefit of the oppressing group did not exist in Cuba’ (1969:140).

Introduction

In the previous chapters. the pronoun er and the class of R-pronouns was introduced at several points. In this chapter I shall focus on the pronoun er mainly in relation to its so-called dummy-pronoun status. The dummy-pronoun appearance of er is only one of the many possible functions of er. I shall argue that in all its uses er is an adverbial pronoun that may enter tnto a variety of syntactic relations. In some sense the spirit of this chapter is similar to that of the preceding one. Het was argued to be a referential NP in all cases. This was seen to be particularly relevant with respect to its so-called dummy use. The analysis of er as an adverbial pronoun is in fact the traditional view. as put forth in the Bech's(1952) article ‘Ueber das niederländische Adverbialpronomen er'. Thi.s analysis differs from most. if not all. recent analyses of er. These generally take the different syntactic functions of er as an indicati.on or motivation for the existence of several distinct pronouns which may vary in their categorial status and inherent syntactic properties. A different view in a completely different framework is expressed in Kirsner (1979). Basing his account on the conception that one form should correspond to one function/meaning. he argues in favour of a unificatory analysis of er that is based on the idea that the meaning of er is constant and can be described as presentational. I agree with Kirsner that an optimal theory of er should relate all occurrences of er. I do not agree with him that such a unificatory approach should necessarily be based on a particular meaning of er that is present in all specific instances. It can be argued that er possesses inherent properties which allow er to appear in a variety of syntactic constructions. Differences between distinct occurrences of er should then be considered to follow from the distinct syntactic relations and not from the appearance of different pronouns which happen to have the same surface realization. Paradigmatical as well as syntagmatical differences can be observed, such as the possibility of substitution of er by other R-pronouns and the maximal domain within which the syntactic relation in which er participates holds.

This book was submitted as a doctoral thesis to the Faculty of Arts of the University of Amsterdam under the supervision of Prof. S.C. Dik and Prof. A.D. Leeman. It owes much to the interest and criticism of a number of friends both in the Institute of Classical Philology and in the Institute of General Linguistics.

Prof. Leeman stimulated the work on this thesis by his continuous interest in both my personal and my working circumstances and greatly contributed to style and content.

Prof. Dik left no paragraph unchanged by his merciless criticism and has made it permanently impossible for me to repay to him what he has done for me in this and former studies.

Prof. Ruijgh did not confine himself to advising the Faculty on this book - actually, that would have been rather hard to imagine - but made many pertinent remarks.

Mrs. A.M. Pool-Bolkestein's critical remarks are almost the size of a book themselves. Miss E. Vester and Drs. A. Rijksbaron, too, expressed their scepticism abundantly. They all spent many hours in discussing problematic sections with me. Drs. J. J. L. Smolenaars proved not to be the layman in linguistics he would make me believe he is. Mr. N.S.H. Smith endeavoured to make the language this study is written in at least look like English.

A number of assistants in the Institute of Classical Philology as well as Mrs. A.M. Pool-Bolkestein, Miss E. Vester, Drs. A. Rijksbaron and Drs. J. J. L. Smolenaars helped in reading the proofs and so did my wife. She has for several years bravely endured the fate of being married to someone writing a thesis. More than to anyone I owe her a debt of gratitude and an apology.

In this chapter I will deal with the closeness of relationship ('affinity’) between constituents. One aspect of this concept is the optionality of some constituents with respect to others. Optionality vs. obligatoriness of constituents bears on the concept of modification. Since the definition of adverbs is formulated in terms of modification, clarification of this notion is required. It will appear that ‘modification’ is not clearly defined. In this chapter I will suggest that adverbs occur with at least three different degrees of affinity towards the verb, in other words that adverbs occur at three different levels in the structure of sentences.

Modification

In the section on syntax of the fourth chapter (4.3.) it was shown, that given the definition of adverbs and given the words that are generally assigned to the category of adverbs, the definition - if conceived of in a generative way - on the one hand allows for constructions which are in fact ungrammatical, on the other hand excludes constructions which are grammatical. The notion ‘modification’ was accepted without a discussion. However, this notion is not at all as clear as its frequency of use might suggest. Therefore, I will now examine it in detail.

Modification is considered to be related to the notion of optionality of a constituent with respect to another constituent, especially the verb. This notion ‘optionality’ is problematic as well. It will appear that the notion of modification is closely connected with the sentence model that is tacitly or not assumed in traditional and modern grammar. I will try to reconstruct the model within the framework of which Kuhner-Stegmann describe ADJUNCTS and compare it with other models

Modification interpreted as optionality with respect to the verb

The notion ‘modify’ (German ‘bestimmen'; French ‘determiner’) plays an important role in Latin and other studies, but is seldom defined (Swiiste 1963: passim;.Crystal 1967: 48). Kiihner-Stegmann are a notable exception. These authors recognize two main types of modification ('Bestimmung’): (a) modification of the subject, (b) modification of the predicate. By the side of modification of the predicate they recognize ‘addition’ ('Erganzung’) to the predicate (see below). The subject modifiers are called attributive.

This chapter will deal with the relationship between certain adverbs and a subclass of so-called coor(Unating conjunctions, which I call connectors!. These connectors are different from both coordinators (ch. 7) and subordinators, which will be dealt with in the next chapter.

Traditionally the category of conjunctions is divided into ‘coordinating’ and ‘subordinating’ conjunctions. German terminology as it appears in Kiihner-Stegmann and Szantyr is rather confusing. The following scheme will perhaps help.

English Kuhner-Stegmann Szantyr

conjunctions Konjunktionen Konjunktionen (= Bindeworter)

coordinating Bindeworter Partikel

subordinating Fiigeworter Konjunktionen

The category of coordinating cdnjunctions is usually subdivided in Latin grammars into four or five subcategories. They are

• (i) copulative, e.g. et ('and’), etiam ('also’), adhuc ('yet’);

• (ii) adversative, e.g. sed ('but’), quidem. ('at least', ‘certainly’), tamen ('still’), ceterum ('but', ‘on the other hand’)

• (iii) disjunctive, e.g. aut ('or’)

• (iv) causal, e.g. nam, enim ('for’), tamen ('still', ‘yet’) in Szantyr (p. 497)

• (v) consecutive (conclusive, illative), e.g. ergo, igitur, itaque ('thus', ‘therefore’)

These subcategories are distinguished on semantic grounds. Syntactically, however, members of several different subcategories can be shown to behave in a comparable way. One subcategory of such semantically different but syntactically similar words is formed by what I call ‘coordinators'. As such I recognized (7.3.2.) et, atque, ac, -que, cum … tum, et . .. et, ut … ita (copulative coordinators), aut, vel, sive . .. sive, -ve (alternative coordinators) and at and sed (adversative coordinators) . It is clear that ‘coordination’ is meant here in a stricter sense than in the traditional term ‘coordinating conjunction'. Traditionally, everything that is not subordination is called coordination.

By connectors I understand roughly all so-called coordinating conjunctions minus coordinators. I do by no means suggest that connectors have very much in common. I am aware, too, that since the term ‘coordinating conjunction’ is vague ‘minus coordinators’ is vague as well. I would certainly reckon among them autem, nam, enim, ergo, igitur, itaque. Characteristics of the latter three will be discussed below. So far the term ‘connectors’ is intended in the first place as a cover term for several subcategories that are themselves negatively defined. A rough definition might run as follows: connectors are those words which establish various semantic relations between paratactic sentences, are no coordinators, and can be shown not to be adverbs either.

Intention

This study deals with a number of aspects of the words which are usually called adverbs in Latin. It contains on the one hand a critical discussion of their treatment in Latin grammatical studies - the characteristics attributed to them, their relationship to other words - and on the other hand a discussion of the conditions that have to be met in order to achieve a better (sub)classification - general problems of classification as well as criteria for effecting such classification - and a better description of the functions of adverbs in larger constructions. The study contains, therefore, both language- specific sections and more general ones. I have tried to write the passages specifically dealing with Latin in such a way that they are clear enough to the non-Latinist, the more general passages in such a way that they are understandable for Latinists who are not acquainted with recent developments in linguistics. I can only hope that I have avoided the risk of being either obscure or trivial to both groups of readers.

Surveyo[the content

The book consists of three parts. The first deals with general problems. The second part (ch. 3-7) contains a discussion of the class of Latin adverbs, the criteria used to define them and a discussion of a new means of achieving a subclassification. In the third part (ch. 8-11) I deal with the relationship between adverbs and other so-called uninflected words with respect to syntactic behaviour. Clearly, a fourth part might have been added dealing with the syntactic characteristics of adverbs and other constituents (such as ablative case forms ‘of nouns, prepositional phrases, subordinate clauses) which are often called ‘adverbial adjuncts', ‘adverbials', etc. Such a study would require to be of considerable length and exceed the bounds of a dissertation. It is my intention, however, to devote a separate monograph to this subject.

The general problems dealt with in part one concern the nature of the material we have to work with (ch. 1) and classificatory problems (ch. 2). I will expound the idea that the work of the Latin linguist is seriously hampered by the fact that it has to be based on a closed corpus, which has been selected to a large extent as a result of literary and other not primarily linguistic factors. On the other hand the linguist's work may be facilitated by his knowledge of general linguistic facts.

In this chapter I will discuss some general problems of classification, viz. the purposes ‘of classification, the types of criteria on which a classification can be based, the relation between these types and, finally, the relative importance of the criteria and the number of criteria required.

Purposes ofclassification

Linguistic items can be grouped into classes on account of certain characteristics they have in common. Word classification, such as the classification of adverbs, is a specific instance of this general problem. The purposes of such a classification are to obtain a higher degree of generalization in the description of the specific language the linguist is working on, to make it more adequate since, while classifying, the linguist is forced to state explicitly which items are similar, and in what respects, which items are different, and what is the relation between the various classes to which the items are assigned. The description becomes simpler since identical or similar phenomena can be described together in one statement (Crystal 1967: 25-7; Van Wijk 1967: 235; Schopf 1963: 63). Classification should, of course, be based on a careful analysis of an adequate sample of material (cf. 1.1.).

Once the linguist has reached a satisfactory classification and has thereby left the stage of discovery he may incorporate the classes in the grammar. The way in which this incorporation is effected will be determined by the type of grammar the linguist has in mind. Traditionally the incorporation took the form of a more or less explicit defmition, like Priscian's adverb definition: ‘the adverb is an invariable part of speech that modifies a verb’ (cf. ch. 3). One may also indicate the membership of classes in rewriting rules, as was usual in earlier versions of transformational generative grammar (Bach 1964: 27-9). For example,

Adverb → nunc ('now’), heri ('yesterday’), etc.

The (sub)classes involved may either be closed and then we can enumerate their members, or open, and then we can only characterize them in some way or other. For example, local adverbs (hic ('here’), etc.) could be enumerated not only because there are not many of them, but also because in Latin there seems to be no way of adding items to this subclass. On the other hand, the subclass of so-called regular adverbs like misere ('miserably’), feliciter ('happily’) is open, since items may be added any time: the formation of these adverbs is productive.

This chapter will be devoted to what Roman grammarians have said about the adverb. It is not my intention to present an exhaustive discussion of the different opinions held by each grammarian, of their sources or mutual relations. Those who are interested may consult Jeep (1893: 268-82). I will give only what seems to be relevant to the modern view on adverbs and confine myself almost entirely to Priscian's Institutiones grammaticae and to Charisius’ Ars grammatical..

The definition of the adverb given in most modem handbooks is based to a large extent on what Priscian, Charisius and others have said about it. In their turn, they were heavily indebted to their Greek predecessors. In fact, the very term ‘adverbium’ is a translation of Greek epirrhema. For the chapter of Priscian, which I take here as a base, one may readily find almost verbatim parallels in e.g. Apollonius Dyscolus. However, I will neglect these.

Priscian and others endeavour to distinguish between the ‘parts of speech’ by giving the semantic aspects of each class (Priscian 2,4,17: ‘proprietates significationum’), following Apollonius (Robins 1951: 64-5; 1967a: 57), but, in fact, most definitions are mixed (Schopf 1963: 64). They contain information about syntax, semantics and morphology, and we find stray remarks about accent as well. I will now first discuss the definitions and dispersed general observations under these headings, next discuss a few specific subclasses and individual adverbs and, finally, add a remark about the primary importance of word-classes in Roman grammar.

General properties ofadverbs

Syntax

The adverb is placed with a verb (Prisc. 2,4,20), ‘more suitably’ (aptius) before it, just as adjectives are placed before substantives (15,6,39; 16,3,16). Adverbs may be postponed, Priscian observes, with the exception of all monosyllables such as non ('not’), ne ('that not’), dum ('while’), temporal cum ('when’), affirmative per ('very’), and vel, when meaning valde ('very’). There are several subclasses, too, that are usually placed before the verb: adverbia demonstrativa like en, ecce ('look’), interrogativa like cur, quare, quamobrem ('why’), adverbia’ hortativa like heia, age ('come on’), adverbia similitudinis like quasi, ceu, veluti ('like', ‘as it were’), adverbia vocandi like heus ('hey! ’), adverbia optandi like 0, utinam ('would that’).

On Latin Adverbs was originally published in 1972. This is an unchanged reprint. Certain topics have continued to interest me, notably the hierarchical structure of sentences and the positions in that hierarchy where adverbs can be found, as well as coordination. I refer those interested in my later writings on these topics to: Harm Pinkster ‘Attitudinal and illocutionary satellites in Latin’, in: Words in Their Places, ed. by H. Aertsen et al., Amsterdam, Free University 2004, pp. 191-198; Harm Pinkster ‘La coordination’, L’information grammaticale 46 (1990), pp. 8-13. Much of the book was used for my Latin Syntax and Semantics (see: http://cybergreek.uchicago.edu/lss/) and will be used in my forthcoming Latin Syntax.

In this chapter I will discuss the question of how we can best account for the fact that the majority of adverbs are related formally and semantically to words belonging to the classes of adjectives, nouns, pronouns and verbs. In principle two solutions are possible:

(1) adverbs are ‘forms of adjectives, nouns, etc. The implication of this solution is that adverbs do not constitute a word-class of their own (are a pseudo-category, as Bergsland (1940: 53) puts it). On ‘form or see 2.2.2. (2) adverbs are related to adjectives, nouns, etc., but cannot be regarded as forms of them.

In the first solution adverbs are inflected forms, in solution (2) they are uninflected or, in other words, invariable. The formal resemblance is based. on a’ derivational relation. If they are regarded as uninflected, it is not immediately clear how comparative and superlative adverbs are to be described. Should we speak about comparative and superlative forms of adverbs or about adverbs related to comparative and superlative adjectives?

In the traditional view it is held that the adverbs meant here are derivationally related to adjectival, nominal, pronominal and verballexemes (or in traditional terminology: are derived from the stems of adjectives, nouns, etc.). It has been suggested, however, from ancient times onwards, that adverbs are not essentially different from inflected forms. Thus the idea that adverbs are words in the ·'adverbial case’ has been maintained by Chrysippus according to Steinthal (1890: I, 302; criticism in Calboli 1971: 115-22). It has received the approval of Hjelmslev (1935: 4).

The discussion on the advantages of the inflectional view over the derivational view has been focused on the relationship between adjectives and adverbs. As such the problem is relevant for modern languages as well.! Bergsland (1940), however, takes adverbs related to the other classes mentioned, into account, too. Three different versions of the inflectional view occur. The first views adverbs and adjectives as forms of the same lexeme, the second regards adverbs as constructed from adjectives by the addition of an empty morpheme, the third regards all four types (adjectival, nominal, etc.) as adverbial case forms.

Adverbs and adjectives are both forms ofone lexeme

In the inflectional view it is held, of course, that adverbs can be accounted for by the constructions which the lexemes they realize enter into.

This chapter will mainly deal with diachronic observations, since the relationship between adverbs and prepositions is seen in most grammars in a historical perspective. Since, however, conclusions are often drawn from this as to their synchronic relationship a discussion seems in place here. First, however, I will make a few remarks about the synchronic relationship itself in so-called classical Latin.

The synchronic point ofview

Synchronically the main difference between prepositions and adverbs is that the former do not occur independently without case forms (Kurytowicz 1949: 131; Q'Brien 1965: 101). We could leave the topic with this observation, if no intermediate type of words existed which share the characteristics of ‘pure’ adverbs and ‘pure’ prepositions. The fact of the existence of such words was mentioned in 3.1.1. with reference to Priscian 15,3,30. Thus, in answer to a question

(1) ubi locum sepulturae dederunt? ('where have they given a place for burial? ’),

we may say intra urbem and in urbe ('inside town', ‘in town’) alike. We might also answer intra ('inside’) alone. *in alone would be ungrammatical. In cases like these intra will be understood with respect to some place known to the speaker (and/or hearer), inside which the burial can take place. Grammatically, however, intra is independent. The semantic dependency is less outspoken in examples like

(2) (animalium) tota illorum ut extra ita intra forma humanae dissimilis est ('just as their outward form is wholly different from that of man, so is their inner nature'; Sen. Dial. 3,3,7),

(extra and intra function as ATTRIBUTE with forma, judging from word order and content). The ungrammaticalness of *in is confirmed by Priscian's words (15,3,30) that e, ex, a, abs, de, pro, sub, in, and others ‘never do occur without caseforms,.

Other words of the intermediate type to which intra belongs probably never occurred independently, but can be recognized without difficulty. An example is ad when meaning ‘about’ in expressions like

(3) occisis ad hominum milibus quattuor ('after about 4000 men had been killed'; Caes. Gal. 2,33,5).

If ad were a preposition there would be an accusative case form governed by it. In the examples, milibus (ablative case) is part of a normal ablative absolute construction. A similar argument holds for occurrences of praeter ('except’), prope ('about’) and iuxta ('alike’). See Kiihner-Stegmann (:1, 575-9).

Morphologically, adverbs resemble words that are usually assigned to other categories such as interjections, prepositions,subordinating and coordinating conjunctions. Their common characteristic is that they are all indeclinable, invariable, or whatever one wants to call it, that is, they do not belong to the inflectional nominal and verbal categories. A common denominator for words belonging to the classes mentioned, that is often found, is’ ‘particle'. In fact, ‘particle’ is as good as any other term as long as it is used consistently in a morphological sense, as it is by Matthews (1965; cf. 2.2.2.). Since in Latin linguistics (but also elsewhere; see e.g. Ruijgh 1971: 99 on Greek), the notion ‘particle’ is often used in syntactic sense as well, or rather in several syntactic senses, I will avoid it and use rather the term ‘invariable’ .

Actually, there are a number of syntactic (as well as semantic) differences between adverbs and other invariables, to which I will now turn. The first and most important distinguishing criterion that I will discuss is the criterion of ‘sentence-valence’ - I borrow the term from Van Wijk (19'67: 240-1). Adverbs are said to be capable of occurring independently, in one-word-sentences, whereas e.g. prepositions are said not to occur in this way. On the other hand, it has .been observed that interjections do not occur other than independently. I will try to show that this criterion is difficult to apply in Latin in the first place and, secondly, intersects with the traditional class of adverbs.

Adverbs as optionally free forms

The only Latin grammar, in which the sentence-valence criterion is used is O'Brien (1965). He observes (p. 39-40) that adverbs optionally occur as free forms, in contradistinction to interjections, which obligatorily occur in that way, and other invariables, which never occur as free forms. Kurytowicz (1964: 21), too, considers the occurrence of traditional adverbs as free forms the primary criterion by which they can be differentiated from prepositions.

Most words which may occur independently are not likely actually to occur in one-word-sentences unless in an appropriate setting of context and/or situation, e.g. only if an appropriate question precedes or, conversely, if an appropriate answer follows. These words may be said to have ‘restricted sentence-valence’ (Van Wijk 1967: 241).

This chapter has to be viewed on the one hand as a supplement to what has been said about connectors (ch. 10), intended to bring this class of words out in full relief. On the other hand, I will contend at the end of this chapter that at least some so-called subordinating conjunctions resemble adverbs, and could be described as relative adverbs, having a function within the clause that is comparable with the function of other adverbs elsewhere. I will use the term ‘subordinator’ instead of ‘subordinating conjunction’ throughout.

As far as connectors are concerned, I have tried to show in ch. 10 that syntactic distinctions between connectors and adverbs have been neglected in favour of semantic similarities. This is done even more so with respect to the relationship between connectors and subordinators. Subordinators, coordinators and connectors have been regarded as one part of speech from antiquity onwards (cf. 3.2.; Kiihner-Holzweissig 1912: 254; Blatt 1952: 33) in spite of the obvious syntactic fact that subordinators can be used in hypotactic constructions only.

Their, in many respects, identical treatment was (and is) believed - with more or less reservation - to find support in two connected considerations, which are both open to criticism and comment: (i) the historical relationship between adverbs, connectors and subordinators; (ii) the supposed comparatively recent emergence of the complex sentence, especially the hypotactic type. It is problably due to the diachronic preoccupation of Latin grammars that the precise. relationship between parataxis and hypotaxis has not been clearly realized. I now want first to give a survey of the various types of relationships that are possible between sentences and between clauses, next discuss the position of most Latin grammars with respect to these facts and, thirdly, add a few remarks of my own about hypotaxis and subordinators.

Types of connection between sentences and clauses

I will first briefly deal with the semantic relationship between (the content of) sentences. Clauses will be discussed in the next paragraph. The semantic relationship between sentences may become clear in three ways, either implicitly from the intrinsic relationship of two or more sentences without connecting words and phrases (1) or explicitly by means of demonstrative (deictic) words (2) or connectors (3). Since the problem does not concern specifically Latin alone, I give English examples. For the explicit connecting devices I will add Latin translations.

In the preceding chapter I endeavoured to show that generally speaking the relationship of adverbs towards other constituents (both in the phrase and, also, .particularly, in the clause or sentence) is much more complicated than the notion ‘modify’ in the traditional definition would suggest. In this chapter I will discuss three criteria by which adverbs can be classified in a dead language even in. those cases where there are no differences with respect to the constituents the adverbs are combined with. For example, in 4.2. a semantic distinction was observed between e.g. vehementer dicere ('to speak impetuously’) and vehementer displicere ('to displease strongly’), manner adverb and intensifier respectively from the semantic point of view. The question was mentioned of whether there are syntactic differences corresponding with the semantic ones. This chapter will be devoted to questions like this. The three criteria discussed are (1) the use of question words; (2) word order; (3) coordination. Coordination will be shown to be most useful.

The use ofquestion words

In the following examples we find several adverbs constructed with adjectives. These adverb-adjective combinations are constructed with a noun. This can be expressed as follows: Noun ((Adverb) Adjective)(cf. 4.2. - the adverbs are italicized)

• (1) (a) laudo hominem valde nobilem

• ('.I praise the very famous man’).

• (b) laudo hominem vehementer nobilem

• ('… the very famous …’).

• (c) laudo hominem sapienter nobilem

• (lit.’ … the wisely famous …'; cf. (11)-(13) in 4.2.).

• (d) laudo hominemfortasse nobilem

• ('I praise the perhaps famous man'; cf. (18) in 4.3.2.).

• (e) laudo hominem illic nobilem

• ('… famous there …'; cf. (26) in 4.3.3.).

• (f) laudo hominem haud dubie nobilem

• (’ … undoubtedly famous …'; cf. (52) in 6.3.).

Intuitively, these adverbs seem to belong to different subclasses though they share the characteristic of being optional with respect to an adjective as their HEAD.

These differences could be brought to light by examining which questions (or question types) these sentences might occur as answer to. Such questions might be

• (1) (a’) quam nobilis est homo quem laudas? “ :: valde nobilis est ('how famous is the man you praise? :: very famous’): intensifier.

• (b’) quam nobilis …? :: vehementer nobilis est ('how famous …? :: very famous’): intensifier.

• (c’) ? quomodo nobilis …? :: sapienter nobilis est ('in what way is the man famous? :: wisely’) : manner.