Dictionaries of Neologisms: a review and proposals for its improvement

: The aim of this paper is to study twelve modern Dictionaries of Neologisms (DN) in six different languages: Catalan, English, French, Galician, Italian and Spanish. I systematically analysed an array of relevant parameters in lexicographical works of this type. First, the two criteria most used by the DNs to identify the eligible lexical units are: the lexicographical (based on the use of a control corpus) and the chronological criterion. Second, I described and analysed in depth the macrostructure (origin of entries, variants, parts-of-speech) and the microstructure (definitions and senses, citations, examples and contexts, dating and thematic fields) of the DNs. From a general perspective, these DNs are grouped into three types: those developed by an author, by publishing houses and by Observatories of Neology – each with different objectives and methodologies. Third, I have tried to summarize the main strengths and weaknesses of these DNs, and I propose a number of contents and measures intended in particular to improve modern DNs. Finally, I conclude that in the 21st century we can indeed speak of a “neological lexicography”.


Introduction
The study of neologisms has aroused a lot of interest in researchers from different language backgrounds and the fact that there is a myriad of lexicographical works in this area attests to this. There are already many compilations in the literature about these studies, therefore this article aims to focus on some of the methodologies and information they provide. Using contrastive comparison, I shall discuss a part of the most important lexicographical and neological features of these works.
Twelve Dictionaries of Neologisms (DN) in six different languages have been selected as samples. I have focused exclusively on modern dictionaries, published from the 1990s onwards, but they are classical in the sense of being in print. Besides, regardless of their size, I only considered monolingual neological compilations drafted from the perspective of language in general, and not from an exclusively terminological perspective. This leaves out those projects developed with increasingly more powerful databases using the web to publish their results, as they need specific studies of their own. perception speakers have of an item's novelty. An up-to-date reflection on definitions of neologism and criteria to identify and measure them can be found in Cabré & Estopà (2009), especially in sections I, II and IX.
Some authors also point out the problems posed by definitions of neologism, which derive from the vagueness of the term and the difficulties to understand exactly what a neologism is, differentiating it clearly from all other phenomena (see Estornell 2009: 19). Sablayrolles (2002) emphasizes the problems derived from the application of a control lexicographical corpus and from the application of the neological perception criterion, and reflects on the "conception large et scalaire de la néologie" [large and gradual conception of neology] and on the construction of what he calls an "échelle de néologicité" [degree of neologicity] (Sablayrolles 2002: 110 and 105-107). Regarding this concept, which was applied, for example, to Galician by Crespo & Gómez (2008), Freixe suggests that neologicity "depèn d'un conjunt complex de factors, que va més enllà de la freqüència i la novetat, i creiem que un d'aquests factors és la raresa, entesa com la qualitat de transgressió de la regla amb què està formada una paraula" [depends on a complex set of factors, beyond frequency and novelty, and we believe one of these factors is some degree of oddity, understood as the capacity to transgress the rule with which a word is formed] (Freixe 2010: 8). In their discussion of the criterion of neologicity, Cabré, Domènech, and Estopà justify the relative character of this concept as such, on account of: It is a well-known fact that novelty does not exist on its own, but only in reference to something else. Therefore, when we talk about neologisms or new words, we must begin by defining for whom a lexical unit is new and in relation to what code (Cabré et al. 2014: 15) For this reason, their working methodology, "in order to establish the neologicity of a word," consists precisely of choosing the "lexicographic criterion" (Cabré et al. 2014: 15).
Finally, some others organize their works according to two other factors: frequency of use and sociopragmatic dissemination. Thus, Kerremans defines neologisms as form-meaning pairings (in one of the three possible combinations), i.e. lexical units, that have been manifested in use and thus are no longer nonce-formations, but have not yet occurred frequently and are not widespread enough in a given period to have become part and parcel of the lexicon of the speech community and the majority of its members (Kerremans 2015: 31-32).
Brought to you by | Universidade de Vigo Authenticated Download Date | 12/12/16 12:47 PM I don't want to conclude this section without recalling, as Boulanger has suggested, that "il appert que la néologie (ou le néologisme) ne devient visible et palpable que dans l'orbite du dictionnaire" [it seems neology (or neologism) only becomes visible and tangible orbiting around the dictionary] (Boulanger 2010: 68). In the following pages, I want to explore precisely this area.

Origin of Entries of DNs
Obvious differences aside, the twelve dictionaries in the sample all show a list of new words or neologisms in alphabetical order and with specific information about them. However, their objectives, regardless of the language, do not always match. In any case, such variation does not hinder a comparative analysis, quite the opposite: it is an undeniable advantage that they have different objectives, as different methodologies contribute to enrich and enlarge the vision that these works on new words offer their users.
First and foremost, I need to analyse the criteria used by each DN to identify the eligible lexical units. There are different criteria that may be used -lexicographical, time span of use, frequency and socio-pragmatic reasons. The first criterion, however, is the one more often present in the dictionaries at hand.
This criterion is based on the use of a control lexicographical corpus. In this way, a word may be included in a DN if its form or meaning is not recorded in general-purpose dictionaries in the control corpus. The dictionaries NQuotidiani, DParaulesNoves, ParoleNuove, NPalabrasGalegas, NDVocesUsoActual and AmongNW all use this criterion. Control corpora may be just a single dictionary (NDVocesUsoActual, or the Spanish Academy of the Language), two (NQuotidiani), three (NPalabrasGalegas, ParoleNuove), four (DParaulesNoves), or nineteen (as for example the AmongNW, where not only seven general dictionaries were used, but also six of British English and six of new words), always in their most updated version at the time when the DN was drafted. In this sense, the AmongNW stands out in the sample, a fact that is, however, justified due the wide time span covered by this dictionary (cf. §4.3). In general, all DNs in the sample share a specific feature -they are the outcome of planned work, usually by the Observatories of Neologisms of each language.
Using the criterion of lexicographical exclusion can lead to a more or less restrictive outcome in two ways: in the efficiency with which the DN incorporates neologisms that meet this criterion, and in the rigour applied to allow for exceptions to this rule. These and other questions are addressed in §3.
Seven DNs in the sample explicitly or implicitly used criteria linked to the concepts of "new" and "newly-coined" applied to the lexis of a language. Therefore, the potential neological interest of a particular word or expression depends in this case on the time parameter above all, i.e., on whether something is new or current, either as the only criterion or combined with others.
The NEspañolActual followed a strict chronological criterion to only include new words documented in the period under review. It was, however, necessary to apply filters for inaccuracies and loanwords.
The DNLenguaEspañola includes some general reflections in its foreword that lead to the conclusion that several criteria were combined: the time aspect, thus selecting new words and new meanings (also colloquialisms or forms that come from specialised language), and the lexicographic perspective as general-purpose (unspecified) dictionaries "por su volumen y complejidad no pueden recoger los vocablos y acepciones nuevos" [because of their volume and complexity cannot include the new words and new senses] (from "Presentación", without page number).
The novelty criteria also underlie the selection of the MotsActualité, which incorporates old words that are used for modernity and that the author wants, in terms of their current use, to gloss or explain. The NW21Century also uses the criterion of newness and it is the author who freely chooses the two hundred neologisms included. In this case, he not only uses a control corpus (in any case only some of his examples are included in dictionaries), but he also tries to underline that the definition of the terms as "real words" does not depend on the fact that these words are recorded in traditional dictionaries while he also claims that their longevity does not depend on lexicography either, but on social use.
Two other DNs (DNouveauFrançais, OxfordDNW) combine time and criteria of neological perception to validate the newness of a term. The DNouveauFrançais, apart from the perspective of the author, also Brought to you by | Universidade de Vigo Authenticated Download Date | 12/12/16 12:47 PM includes other contributors. Furthermore, before the team incorporates new forms, they let two years go by to "vérifier qu'il s'était installé" [confirm that it is consolidated] and so that "chaque mot a été passé au tamis du temps et des usages" [each word has passed the test of time and use] (p. 6). From a lexicographic perspective, this DN highlights that, although some of these words were already documented in general French language dictionaries, this was done through very brief or obsolete definitions that need updating.
Regarding the OxfordDNW, selection in this case focuses on new words, sentences or meanings popularised in English during the 1980s and beginning of the 1990s. Nevertheless, the ones included are not only those created during that period, but also those that people felt were new, to the extent that "the deciding factor has been whether or not the general public was made aware of the word or sense during the eighties and early nineties" (p. v). There is an underlying acknowledgement of the fact that these words are subject to fads and that some could appear in the language before the period analysed. There is also an explicit reference to the fact that new words were not selected on the basis of their inclusion in general dictionaries and thus a fourth of the OxfordDNW entries are documented in other lexicons.

Macrostructures
There is great variation in the number of entries in the DNs in the sample: three of them include a few hundred forms (one hundred in the MotsActualité -although the figure triples if we consider all the words that are commented upon but that do not appear as lemmas but only in the index -two hundred in the NW21Century and four hundred in the DNouveauFrançais), three are around -always above -a thousand entries (DParaulesNoves, NEspañolActual, NPalabrasGalegas), four are between two thousand and five thousand (AmongNW, OxfordDNW, NQuotidiani, ParoleNuove) and, finally, there are three that are close to ten thousand (DNLenguaEspañola), or well above that (NDVocesUsoActual). Table 1 shows the number of entries, rounding off when the work does not include this detail in particular: I can anticipate that the remarkable final quality of any DN does not depend on the higher or lower number of entries. What is important is that each accomplishes the objectives for which it was written and that it is done from a vision of scientific rigour, the same as for any other discipline in linguistics. As I mentioned before, those dictionaries using a control lexicographical corpus may do so either very rigorously or in a lax way. Any compilation of new words that spans a long period of time without being published, runs the risk of seeing some of these words recorded in new editions of general language dictionaries (we should not forget that precisely one of the objectives of neological studies is to offer potential candidates to update general purpose dictionaries). If this happens, the DN can either (a) eliminate all these neologisms that "lost" their lexicographic novelty while they were being drafted or (b) use some of them.
The AmongNW is an example of the first option: the fact that its content was published as subsequent contributions from the 40s onwards, approximately on a half-yearly basis, allowed the authors to strictly Brought to you by | Universidade de Vigo Authenticated Download Date | 12/12/16 12:47 PM stick to this principle. The NDVocesUsoActual also applied this criterion in a strict way and the authors scrupulously eliminated all the entries or meanings in the latest version of the dictionary that had been recorded in their control corpus. However, they kept all the entries or meanings documented in the study period if they had already been included in the first edition of the NDVocesUsoActual, published almost a decade before (they specify this in a note; these forms add up to almost one third of the total).
The DNs that followed the second option usually introduce a note in the microstructure stating that a particular neologism was already documented in general dictionaries. In the ParoleNuove, these forms make up 5% of the total, with 17% for the NPalabrasGalegas and 28% for the DParaulesNoves (these percentages include new forms with some graphic variants; the figures for the DParaulesNoves do not include 12% of forms present in general dictionaries but with a neology label). In the NPalabrasGalegas these examples are justified because they were in their databases and they therefore became "mostra dun uso que tardou algún tempo en aparecer nos dicionarios galegos" [a sign that it took some time before Galician dictionaries included this use of the word in question] (p. 8).
On the other hand, if the criterion of lexical exclusion were applied on direct documentation, the DNs could include the highest number possible of neologisms or, on the contrary, apply a filter and publish just a proportion of them, but not all. This second option was taken by the DParaulesNoves, which compiled less than 5% of the total neologisms registered in their Observatory of Neology and the NPalabrasGalegas, and published a little bit more than 9% of the available new forms. The selection criteria were, in both cases, frequency (the most widely documented) and significance (presence over a longer period of time). The NPalabrasGalegas also considered that significance can be found in spontaneous neologisms in the corpus under study.

Origin of Entries
Good practices in most DNs in this study include identifying the source of the neologism. Obviously, this aspect is supported by the quotations used (cf. §4.2). There is a common nexus in all DNs: using the language of the media as direct documentation. This, in fact, shows how the media play an increasingly important role in our society and the role it seems to play in informing our language. Quotations do not only include this area, or a subtype, they also include the internet in its public-institutional or private side, and speeches.
The ParoleNuove, DParaulesNoves, NQuotidiani and NDVocesUsoActual used the media; the NPalabrasGalegas and NW21Century, the media and internet, and NEspañolActual, which also included different types of books. The MotsActualité was based on the language of media and politics; the OxfordDNW focused on media, works of fiction and general magazines of different English language countries while the DNouveauFrançais, interested in language in use, studied the French language used on the radio, television and internet, but also language used in street conversation and businesses. The AmongNW shifted from the written press to magazines and books, but also used manuscripts and speeches. Additionally, I presume, as this is not specified, that the DNLenguaEspañola used sources from the media, some specialised domains for new terms related to computer science or medicine, for example, and colloquial speech.

Variant Treatment
New forms in any language, as novel forms with (very) low or inexistent presence in dictionaries, are particularly sensitive to formal variation. The DNs in the sample tend to be very receptive to incorporating information about this through different strategies. There are three general strategies: including variants as independent entries; clustering them under a single lemma or including them in the microstructure. However, the former division is not to be seen as an exact strategy but as a way to understand the different DNs from an organisational perspective. There are usually different treatments in each DN, as we will see next.
Brought to you by | Universidade de Vigo Authenticated Download Date | 12/12/16 12:47 PM Three dictionaries tend to use the first strategy: The DParaulesNoves includes all variants as entries. Only one is defined, and although the criteria for selection are not specified (sometimes because of the higher or lower presence of examples, we could come to the conclusion that frequency of occurrence was used), all entries include examples and internal cross-references at the end of each article. For example, only alto-el-foc is defined, with alto el foc and altoelfoc.
Variants also have an independent entry in the NPalabrasGalegas, without a definition, but with examples and cross-referencing back to the main entry (where variants are also mentioned). In the introduction of the work, frequency is listed as the main selection criterion for the main entry. Therefore, ranking is the main entry with several variants, ránking and ranquin.
In the NDVocesUsoActual variants have their own entry and offer the same information as the main lemma, except for the definition, as the usual practice is to refer back to it (for example, s. v. rottweiler includes a definition, while rottweiller just refers back to the former). Nevertheless, sometimes there are definitions for all variants: in acuaplaning, after referring back to aquaplaning, the definition is included.
Six dictionaries chose the second strategy: The NEspañolActual includes all variants in the same entry (for example, emoticón or emoticono), in bold type and without any orthotypographical differences between them. In some cases, one of the variants is referred to as less frequent (for example, "asinoterapia o, menos frec., asnoterapia").
In the NW21Century, variants are shown in the entry, to the right of the lemma, with "also" before them, printed in bold but in a smaller font. Thus, the lemma dotbam is followed by the variants dot-bam and dot bam, or marmalade dropper by muffin choker (cf. Appendix X).
In the NQuotidiani and ParoleNuove, variants appear in the entry, in parentheses, after the lemma chosen for the article (in Italics or regular font for the first one, and in regular font for the second one, but never in bold). For example, the lemma sentenza-choc includes the variants sentenza choc and senteza shock (NQuotidiani) and for autorganizzato there are two variants auto-organizzato and auto organizzato (ParoleNuove).
The DNLenguaEspañola includes variants just after the entry, without any typographical markers, preceded by "also" and, if there is any information on the origin of the word, after that. For example, for futbito the variants fubito and fulbito are included.
The OxfordDNW places the variants in the entry, to the right of the lemma and after grammatical information. They are usually preceded by "also written" (also "sometimes", "frequently", etc.), in bold the same as the lemma, but in a smaller font (cf. Appendix XI). For example, for fly-tipping, variants included are: fly tipping and flytipping.
The third strategy is present in the MotsActualité, including all cases of variants in the explanatory article of each of the entries, without referring to the independent lemmas. Thus, for the lemma flexisécurité the article claims that there are other variants such as flexicurité, flexécurité or flex-sécurité (cf. Appendix V).
Two DN used strategies that mix these last two: The case of the DNouveauFrançais shows variants either by the entry (in parentheses but with identical typography: for example, swag or swagg), or else inside the article under the heading "variants" (also "equivalent"). This information is selectively offered, because in some cases some of the variants included in the quotations, for example, alzheimer are not presented as variants of the new word (Alzheimer).
The AmongNW is a special case as the dictionary is divided into two sections (apart from the introduction), and only the first part is relevant for my study at this point. This part includes a list of new forms operating as lemmas in alphabetical order (there are some cross-references too). Variants are located in two spots: either by the lemma, with the same typography and separated just by commas, or else inside the articles, in smaller font, or by other new forms related to the lemma (also marked, all of them, in bold and separated again by commas). An example is intifada and intifadah (cf. Appendix I).
Some  Looking at the figure above, one can see that there are three main types of DNs regarding the number of variants against the total entries of each dictionary: those with very low percentages, between 1 and 3% (NPalabrasGalegas, NEspañolActual), those around 10% (DNLenguaEspañola, AmongNW, NDVocesUsoActual, DNouveauFrançais, DParaulesNoves, OxfordDNW, ParoleNuove) and those showing greater permeability to the introduction of different variants (NW21Century, NQuotidiani). In dictionaries belonging to big publishing houses with large lexicographic collections, there is relative homogeneity (DNLenguaEspañola, OxfordDNW), except for the NEspañolActual. Likewise, DNs under a particular Observatory of Neologisms show certain parallels (DParaulesNoves, NDVocesUsoActual, AmongNW, ParoleNuove) while the NPalabrasGalegas breaks the trend with an acute scarcity of variants, and the NQuotidiani with its abundance, almost doubling the average (this disparity is remarkable considering that the NQuotidiani and ParoleNuove share both the language and the working methodology). Is there a reason for this disparity? Were the neologisms of the Italian press in the 1998-2003 period (NQuotidiani) (ortho)graphically more unstable than those from the 2003-2005 period (ParoleNuove)? Without discarding this possibility, there could at least partly, be other hypotheses to explain such striking disparity. On the one hand, the number of examples per entry is similar (cf. §4.2), although the figure is slightly higher for NQuotidiani (2.12) than for ParoleNuove (1.81), which translates into a higher probability of variation for the former. On the other hand, the number of idiomatic expressions in NQuotidiani is almost seven points higher than in ParoleNuove. Besides, the varieties of idiomatic expressions in NQuotidiani almost double those of ParoleNuove. It should not be forgotten that NQuotidiani includes many more loanwords that ParoleNuove.
Author-based dictionaries show that while the DNouveauFrançais falls within the average, the NW21Century is especially prone to including variants. In the latter case, it is true that many variants are offered (for example, Wi-Fi, Wifi, Wi-fi and wifi, s. v. WiFi), but no examples are given of them. In this dictionary, there is a change of stem in one in every seven variations (as Delia effect and Delia power), and there are also some cases of diatopic variation (for example, marmalade dropper and muffin choker; cf. Appendix X).

Analysis of Entries per Part-of-Speech
The first aspect be considered is this regard is that all DNs revolve around lexical categories almost exclusively. Within that, it is mainly nouns and adjectives, respectively, that are most widely represented (for NPalabrasGalegas and ParoleNuove see Rodríguez Guerra 2015, 402). This may be due to the fact that they show greater formal diversity and include in particular, derived forms -with prefixes, for example, outyear (AmongNW), pre-estrena [preview]   This does not happen with adverbs. Though most DNs include them (there are none in the NW21Century or DNouveauFrançais for example), the fact is that the proportion is always small: 0.1% in the ParoleNuove and in the DParaulesNoves, 0.2% in the NPalabrasGalegas, 0.3% in the OxfordDNW, 0.4% in the NEspañolActual, 0.8% in the AmongNW (for the period 1986-1991) and the DNLenguaEspañola, and 1% in the MotsActualité. Besides, most of the time it is not simple adverbs that are included, (2a), but compound adverbs; that is, adverbial phrases (2b). Such low figures are better understood if we consider that all adverbs ending in -mente in Romance languages are not usually included in the DNs, the examples in (2a) are an exception, a fact explicitly explained in the NDVocesUsoActual (p. VIII, n. 5). In the DNs' entries, we also find "espressioni polirematiche" (NQuotidiani and ParoleNuove) or "expresiones pluriverbales" (NEspañolActual) in an important proportion (31.8% of the entries in the NQuotidiani and 28.7% in the ParoleNuove). For complex units, there is sometimes a generic denomination, either as phrases (without any further specification), or as expressions. Thus, for example, the DNP only speaks of phrases, and they include examples such as a l'americana, at a phrase level (although there are exceptions, see below). The NDVocesUsoActual also includes units seen only as phrases, for example, déjà vu. On the other hand, some entries in the NEspañolActual are just identified as "expressions", and there are some Latin ones such as bonus malus, but in general they tend to be expressions from English such as community manager.
Specific mention must be made of complex neological units at sentence level. They are not present in the NPalabrasGalegas, but they are present in other dictionaries, although mostly with a very small sample. The DNP only includes one example of this kind as a subentry (see the previous paragraph) of canya, which is identified as a phrase. There are some exceptional examples in the NQuotidiani, such as non fa una piega (analysed as an invariable noun phrase) or fare flop and dire cose di sinistra (both identified as a verb phrases), and similarly in the ParoleNuove for cases such as guarda e compra (invariable noun phrase) or fare sistema and fare squadra (verb phrases). The AmongNW includes some subentries such as when you get down to the bottom line (of 1973) and the NW21Century only includes jumps the couch as "idiom". Entries in the OxfordDNW also include some examples of verb phrases such as chase the dragon and others such as have (got) a lot of bottle inside the articles. The MotsActualité includes three examples of such sentences under the name "formulae", for example, caracoler en tête des sondages [to be riding high in the polls]. The NDVocesUsoActual includes the entry by the lemma of the verb, identified as (in)transitive, and then the rest of the structure (with the same font, the initial verb in bold), for example, "aguantar tr. -el tirón" [to weather the storm] (p. 19).
The NEspañolActual includes information about parts-of-speech, and it records some ten expressions such as "muebles, salvar los ̵ ̴ " [to save face] and "piscina, tirarse [o lanzarse] a la ̴ " [to take the plunge]. All these examples are anecdotal and have a large variation in terms of denominations, to the point of categorisation being totally absent in some DNs. In fact, I can conclude that this level of complex linguistic units is not usually found in DNs. There is, however, one exception, the DNLenguaEspañola. There is an important number of sentence-based entries in this dictionary. The verb is the lemma and then the rest of the elements are presented afterwards. All these cases are identified in the verbal category, except for ir detrás de, verb phrase, and ¡ahueca el ala!, interjection. In order to assess their performance, we have to highlight that there are verbs showing many examples: twenty-eight with ir (ir de parranda), twenty-five with tener (tener madera), seventeen with caer(se) and dar(se) (caerse con todo el equipo; dar cancha) or fourteen with pasar(se) (pasarse de listo), to name a few.
The DNouveauFrançais deserves special attention, as it goes beyond alphabetical order into the symbolic: the first four entries are a series of symbols, such as "<3", symbolically imitating a heart, and used to show love for somebody. It also includes conventional expressions such as AAA (present in English in the OxfordDNW, although with different meanings), acronyms (such as A.I.P.) and more complex expressions such as ah, ouais quand même!, or the attributive je ne suis pas amoureux.
Brought to you by | Universidade de Vigo Authenticated Download Date | 12/12/16 12:47 PM Finally, some classical compounds are formed with word-forming elements, prefixes and suffixes, that also have this privileged position in some of the analysed DNs. The NEspañolActual introduces Classical Latin and Ancient Greek word-forming elements with a full microstructure (mainly as the first element: ciber-, macro-...; but frequently as the second, such as -logo). However, there are two particular DNs of the English language that make the best of the potential of these forms. Thus, the OxfordDNW, includes word-forming elements in the entry with a full microstructure: combining forms (such as the first, bio-, cryo-, tele-..., or second elements, -Aid, -babble, -free...) and prefixes (hyper-, pro-...). The AmongNW also compiles word-forming elements as separate entries, either as combining forms such as prefixes B, big or sky, or as suffixes borne or hop, or in any position such as type.

Microstructure
A DN may be characterised by the new forms selected (cf. §3), or by the information on each of them. In my review of these twelve DNs, I now analyse the most important content in each of them. Apart from that, some reflections on the most relevant aspects of such works -in my view -will be included. Other elements, already included in the section devoted to macrostructure, shall not be addressed again (see for example information on parts-of-speech).

Definitions and Meanings
One of the obvious key aspects of all DNs is explaining the meaning of the new words. Although they all show high lexicographic quality in the definitions (see Porto Dapena 2014), the behaviour of each dictionary is unique in this regard. Most of them use lexicographic definitions of a conceptual kind (usually analytical but, sometimes, based on synonymy) or explanations: this is the approach of the DNs stemming from wider lexicographical projects of large publishing houses (OxfordDNW, NEspañolActual, DNLenguaEspañola), as well as those originating in Observatories of Neologisms (DParaulesNoves, NPalabrasGalegas, AmongNW, NQuotidiani, ParoleNuove, NDVocesUsoActual). Some examples of analytical definitions are 'piccola attenzione rivolta a sé stessi per gratificarsi' [a little indulgence to reward oneself] (ParoleNuove, s. v. autococcola) or 'que consume alimentos crudos, principalmente de origen vegetal' [someone who eats raw food, especially vegetables] (NEspañolActual, s. v. crudívoro); synonymous definitions include 'diversió, animació' [amusement, bustle] (DParaulesNoves, s. v. marxa) or 'audaz, dinámico, emprendedor' [bold, dynamic, enterprising] (NDVocesUsoActual, s. v. agresivo); while explanatory definitions, are used mainly for more complex units: 'expression familière que l'on utilise dès que l'on estime que les limites ont été franchies' [informal expression used to imply that the limits have been crossed] (DNouveauFrançais, s. v. abusé (c'est)) or 'se usa para ofrecerle a otra persona su protección' [it is used to offer someone protection] (DNLenguaEspañola, s. v. tener en memoria). Whenever synonymy is used, two or more synonyms are mentioned. Occasionally, however, when new forms already have a perfect equivalent in standard lexicons, we may find a single synonym without a real definition: for example, acúmulo is defined as 'acumulación' (NDVocesUsoActual) and in NEspañolActual enter as 'retorno', farlopa or farla (

characteristics of power and motor torque of an engine equipping a given vehicle] (DFL, s. v. motorisation).
It should be remembered that motor, in addition, has some other acceptations that motorització can hardly take up, for example, 'persona que incita a obrar' [someone who encourages others to act] (DIEC2, s. v. motor, fourth acceptation).
Quite cogently, especially to save space, there are cross-references in the definitions: for example, in DParaulesNoves, the entries cap afaitat, cap rapat and skin head are defined as 'skin' and skin is defined as 'membre d'una tribu urbana d'ideologia nazi' [urban tribe member of Nazi ideology]; or in NEspañolActual macroproceso is defined as 'macrojuicio' and the entry macrojuicio is defined as 'juicio en el que se implica a un número elevado de personas' [a trial in which a large number of people are charged]. This is quite common in the case of equivalent words remix > remezcla or tupper>tupperware>táper, or in the case of those belonging to the same word family, like discapacitante>discapacitar, enrocamiento>enrocarse or vigoréxico>vigorexia (NEspañolActual).
Definitions in DNs rarely use markers to identify the collocations or context. Thus, the only DN systematically including information on collocations is the NDVocesUsoActual, which, for example, defines arrebotijar as 'recoger [algo para meterlo en un lugar]' [pick up {something to place it somewhere}].
Otherwise, the non-conceptual part of the definition may be found outside it, isolated at the beginning or end. There are DNs that use formulae to restrict the use of cross-referencing, either by placing these references at the beginning, separated from the definition by punctuation, or in parentheses: 'of a person, social group, etc.: fully informed about current issues of concern in a particular field' (OxfordDNW, s. v. aware), 'Sur le réseau social Facebook, faux-ami qui signifie contact' [On the Facebook social network, false friend meaning contact] (DNouveauFrançais, s. v. ami) or '(in army use) bemused from living in cactus patches' (AmongNW, s. v. cactus happy). Others try to place them after the definition and a full-stop or a comma: 'Conciliación de lo global y lo local. Se usa principalmente referido a la economía' [Reconciliation of the global and the local. It is used mainly with regard to the economy] (NEspañolActual, s. v. glocalización) or '2. Resultado de salida de datos o información, especialmente en informática' [2. Result of transferring data or information, especially in computing] (DNLenguaEspañola, s. v. output).
The DNouveauFrançais, MotsActualité and NW21Century show a totally different behaviour. As was shown above, these are the three DNs with the lowest number of entries, but these small but well-stocked dictionaries focus all their efforts in offering elaborate articles for each entry. Aesthetics is also a key goal for these dictionaries. First, the DNouveauFrançais combines a lexicographic definition, conceptual, analytical or explanatory, just underneath the entry, centred and in small caps, with the body of the article including reflections or comments to understand the intra-history of each new form. Of all the DNs in the sample, the MotsActualité is, the one that offers more space to each entry: one to three pages. There is a detailed explanation, always from different perspectives (historical, linguistic and cultural), to fully explain not just the meaning of each new form, or the new meanings of existing forms, but also the lexical family of the entry. In this sense, we need to remember that these articles, with many modifications, come from radio chronicles and have an encyclopaedic character close to the original orality. Finally, the NW21Century always shows a full paragraph of six to thirteen lines to explain and contextualise the use and meaning of each entry (this is done without using a traditional lexicographic definition) with concrete reference to the use of the new form.
I conclude this section with a reflection on the meaning of new words. First, one has to emphasize that in most cases these neologisms only have a single meaning or acceptation. This means that the number of entries and meanings are almost the same. Per dictionary, the approximate figures are: 1.02 (AmongNW), 1.05 (NPalabrasGalegas), 1.06 (NEspañolActual), 1.07 (NQuotidiani), 1.09 (NDVocesUsoActual), 1.09 (ParoleNuove), 1.12 (DNLenguaEspañola) and 1.15 (DParaulesNoves). It is clear that the figures are similar and, in fact, six of them show a variation of only 0.07 points. Besides, in the few cases where there is more than one definition per word, the number is usually two or, exceptionally, three. In this sense, the DNLenguaEspañola is an exception, because it has forms with up to six acceptations (for example, enrollar).
Some DNs use the figurative categories or meanings by extension to justify polysemic new words. The DNs where this presence is higher are OxfordDNW, DNouveauFrançais and DParaulesNoves (between 5% and 7%). Ranked below the former, we find the NPalabrasGalegas (1%), NEspañolActual and NW21Century Brought to you by | Universidade de Vigo Authenticated Download Date | 12/12/16 12:47 PM (both with 0.5%), and with still lower percentages, the AmongNW, NDVocesUsoActual and ParoleNuove. Examples of this are badder, including, by extension, the definition 'broyer du noir, quel qu'en soit le motif' (DNouveauFrançais) or escudería 'calquera grupo de persoas que traballan para un mesmo fin' ['any group of people who work in pursuit of the same aim'] (NPalabrasGalegas), and of the noun cascade, where the figurative use is explained as 'in which the information is seen as falling and spreading like a waterfall' (OxfordDNW) or flip-flop 'tentennante, che non assume una posizione decisa' (ParoleNuove). In DParaulesNoves I have noticed a procedure (which is made explicit in the dictionary itself, p. 14) that should be revised: three out of four figurative uses and almost half of the meanings by extension are not defined (as with for example, the second acceptations given for boutique and grandeur for the uses by extension, or with those of doblet and software for figurative uses).

Citations / Examples / Contexts
Any DN is deemed to include quotations to show and prove the existence of the new forms or meanings. They play a fundamental role (and also help in dating the form) and this may be the reason why most DNs live up to the expectation and include them. There is only one, the DNLenguaEspañola, that does not show its sources clearly, therefore examples seem to be ad hoc. All other DNs show real quotations, with complete reference to the date and source where they were documented. Looking at the number of quotations included in each DN and the number of entries, the MotsActualité shows a very low proportion (0.04%), maybe due not only to the oral origin of its articles, but also to the encyclopaedic explanations included, and the fact that examples in them tend to be short (for example, from headlines). After that, we find the DNLenguaEspañola, with 0.6 quotations per entry. There is a series of three DNs with almost total balance between the number of entries and quotations, one per entry (NEspañolActual, NW21Century, OxfordDNW). With a quotation and a half per new form, we find the AmongNW (in the period 1986-1991), four dictionaries are around two per entry (DNouveauFrançais, NDVocesUsoActual, NQuotidiani, ParoleNuove) and finally the two DNs that have more quotations per entry than that, around three and a half, are the NPalabrasGalegas and DParaulesNoves. This relative abundance of quotations, combined with good selection, helps verify and enrich the use of new words in real language use.

Dating
Indeed, if there is a parameter that is second nature to neologisms, it is the chronological parameter and the value of "novelty" it brings with it. This "novelty" factor is diluted and may be lost, or there are more chances of this happening, as years go by.
What period did the DNs cover? First, one has to highlight that most dictionaries make it possible to state ante quem and post quem dates for their collection work, except for the DNLenguaEspañola. After these two, and leaving aside the specificity of the AmongNW, the widest period found is that covered by the NEspañolActual, over twenty-four years. This dictionary, based on a chronological criterion, emphasizes this detail as the "novelty" of neologisms can be seen within a time span of almost a quarter of a century. Periods tend to be far shorter in other dictionaries. Thus, while the OxfordDNW and NDVocesUsoActual studied neologisms for twelve years, seven others worked with periods shorter than a decade: The MotsActualité and DNouveauFrançais analysed nine years each, the DParaulesNoves and NPalabrasGalegas, eight each, the NW21Century seven, the NQuotidiani five and finally the ParoleNuove, three.
The AmongNW deserves special attention. It was initially published as a series of half-yearly articles from 1941 to 1991. This is a work by several authors, but it has been rigorously conducted and properly coordinated to apply the same criteria for five decades, which allows us to analyse data diachronically. In the book version, not only the original articles were included, but also a comprehensive introductory study, a list of the neologisms in order, with their meanings and dates.
Brought to you by | Universidade de Vigo Authenticated Download Date | 12/12/16 12:47 PM It should be noted that, apart from the dates of the oldest quotations, there are DNs that explicitly offer information about the first appearance of a particular new form (using different corpora and textual databases). This is the procedure, for example, used by the ParoleNuove, NPalabrasGalegas, AmongNW or MotsActualité (see Rodríguez Guerra 2015: 403-405 for the first two).

Thematic Fields
Only a few DNs offer information about whether the new forms are linked to any thematic field. This information could be useful in two ways: to recognise some terms that belong to specialised language and are incorporated into the common language, and understand what thematic fields yield more neologisms. There are some exceptional examples in the NDVocesUsoActual (0.03% of senses), such as manolera (tauromachy) or definir (sports).
The NEspañolActual works in thirteen fields and the percentage of classified forms is 19%. The fields that yield more terms are computer science and, far behind, medicine. The DNLenguaEspañola includes over eighty different fields, and although most do not yield many terms, this tagging proves useful as it covers 56% of senses. The most important fields in descending order are medicine, computer science, sports and politics. Regarding the OxfordDNW, eleven important thematic fields are characterised with icons in all the dictionary entries. These fields include science and technology, health and medicine, lifestyle, leisure and sport, people and society, music and politics (cf. figure 2).

Miscellaneous
Apart from all the above-mentioned items, the microstructure of some DNs includes other details. One of the more relevant aspects has to do with the inclusion of usage labels regarding the register or character of the new form. The dictionaries where this can be seen use them in different ways and proportions, taking into account that the total number of definitions involved does not even reach 1% in some (NDVocesUsoActual, AmongNW), while others are between 3% to 5% (OxfordDNW, DParaulesNoves, NEspañolActual, NQuotidiani) and, finally two DNs, the DNLenguaEspañola and ParoleNuove, register the highest number of labels, around 10%.
Regarding specific labels, the most habitual ones are "pejorative" or "demeaning" (ParoleNuove, NQuotidiani, DNLenguaEspañola, NDVocesUsoActual, NEspañolActual), "ironic" (NQuotidiani, ParoleNuove, NDVocesUsoActual), "humorous" (NQuotidiani, ParoleNuove, NEspañolActual, NDVocesUsoActual, OxfordDNW), "colloquial" (DNLenguaEspañola, NDVocesUsoActual), "familiar" (DNLenguaEspañola, NDVocesUsoActual), "informal" (NEspañolActual), "euphemistic" (AmongNW, OxfordDNW), "vulgar" (DNLenguaEspañola, NDVocesUsoActual), "burlesque" (NDVocesUsoActual), "learned" (DNLenguaEspañola) or "formal" (DNLenguaEspañola) -series (5). The DParaulesNoves claims that neologisms are pragmatically characterised by a low level of formality (liquidar, xoriço). However, each dictionary allocates the previous labels unequally, but always with clear preference for a particular term over the others: "ironic" in two Italian DNs, "informal" in the NEspañolActual, "colloquial" in the DNLenguaEspañola, "euphemism" for the British AmongNW and OxfordDNW, and "burlesque" in the NDVocesUsoActual. In comparison to all others, there is a surprising number of vulgar forms introduced in the DNLenguaEspañola (almost one fifth of the labels in this dictionary). This information is very useful, but difficult to implement: for example, paja mental or peque without a specific label in the NDVocesUsoActual; peque in the DNLenguaEspañola presents the label "familiar". Only some DNs include phonetic information. It is only the two English language works that systematically transcribe pronunciation of all neologisms: the NW21Century and OxfordDNW (it is a very common practice in English lexicography in general). Other DNs offer this information on a selective basis, thus for the French DNouveauFrançais, it indicates the pronunciation in some forms such as "2.0" (symbols, signs and acronyms) and the NEspañolActual transcribes some specific cases of loanwords.
The three English language DNs are not averse to the diatopic variation of neologisms. The NW21Century shows most examples, over 95% of them, without any geographic characterisation, but whenever this is relevant, they indicate whether such new forms are from Australian, British or American English. The AmongNW, while it evolved from the American Speech Journal, handled sources from different geographical areas and marked several new forms as "British". The OxfordDNW also indicates whether some of the examples are British or American. In Spanish, the NDVocesUsoActual and NEspañolActual included the American press, but the only geographical information is the one derived from the quotations. As Cabré, Domènech, Estopà (2014) showed for the specific case of Catalan, information on diatopic variation is a Brought to you by | Universidade de Vigo Authenticated Download Date | 12/12/16 12:47 PM good perspective for the analysis of new words, as this can influence the development of dictionaries in the future.
Some DNs include information on synonyms, for example, the DParaulesNoves, AmongNW, DNouveauFrançais or NPalabrasGalegas, but it is the NEspañolActual and, above all, the DNLenguaEspañola where this is done more systematically (over 7% in the former and 13% in the later). The AmongNW and DNouveauFrançais exceptionally add some antonyms.
The DParaulesNoves incorporates, on the one hand, concrete data on the typographic labels of neologisms in the documented examples, while on the other, it indicates whether this new form has been created by an author -over half of the forms fall under this category, excluding quotations from the written press. Finally, the DNs rarely offer etymological information about the new forms: the NEspañolActual explains the origins of some new words, especially classical compounds, abbreviations, acronyms, or foreign words.
From a neological perspective, those DNs that offer systematic information on word formation processes and their elements, such as the NQuotidiani, ParoleNuove and OxfordDNW, prove more modern and useful.

DNs of the 20th and 21st Centuries
After examining the main characteristics of the twelve chosen dictionaries of neologisms, I will reflect critically on their main strong and weak points. After this, I will offer my own suggestions regarding the information an ideal DN should offer its readers now that the dawn of the third decade of the 21st century is fast approaching.

What Our DNs Are Like...
With an average of 75 neologisms per year, AmongNW is one of my favourite DNs, but searching is not always easy because of its compilation format, even if its 'Index of new words with glosses' is an improvement in this respect. Full information about entries may be scattered throughout different chapters of the book's second part, and it is not always easy to find (cf. Appendix I for reality and its word family or intifada). Its control corpus is very large and it also includes some new word glossaries that, being suitable for a periodic publication format, seem less so for a dictionary aspiring to compile neologisms over a longer span of years. Two characteristics must me emphasized -homogeneity in the application of its methodology and the possibility it offers of observing neological changes from a diachronic perspective.
OxfordDNW is a quite comprehensive DN -with its 167 neologisms per year, it is close to the global average (200 neologisms per year). Its weak points lie with the selection criterion (neological perception) and with the length of the period under study (12 years), as a result of which 25% of the new words are already included in general-purpose dictionaries. On the plus side are the treatment of diatopic variation, thematic fields, data on word-forming elements and the almost encyclopaedic information on neologisms and their uses, all of which are quite sound.
In pocket DNs like NW21Century, DNouveauFrançais and MotsActualité, one can notice from the very first page that the criterion used for the selection of neologisms is the subjective preferences of their respective authors. This is good in the sense that the resulting dictionaries are more homogenous in general, but sometimes results in combining some judicious selections with others that seem arbitrary, globally giving rise to unbalanced final products. The information offered by NW21Century, with 30 neologisms per year, is rigorous, but somehow limited. The explanations and glosses given allow readers to understand the meaning of entries, but a lexicographic definition for each would be helpful. Only one example is given for each neologism and the data provided are insufficient to assess the longevity of these "real words". DNouveauFrançais, with 44 neologisms per year, has a remarkable structure that is visually very appealing. Apart from its brief lexicographical definitions, its articles stand out for including explanations about the entry's origin from an etymological or historical perspective, context or use, nuances in meanings, Brought to you by | Universidade de Vigo Authenticated Download Date | 12/12/16 12:47 PM and comparison with traditional forms. The choice of neologisms gives priority, perhaps excessively, to loanwords from English, which are by far the largest in number (possibly due to the fact that the corpus is constituted by oral and social network conversations). The inclusion of some symbolic neologism can only be considered anecdotal. Among the DNs analysed here, MotsActualité contains the lowest number of neologisms per year (11). From the very first article, one notices that this is not a typical DN. It displays an austere typography: entries are followed by all the information which is divided only in equable paragraphs (without category details or lexicographical definitions, but including lengthy encyclopaedic explanations). The target audience of this DN seem to be listeners, rather than readers.
DNLenguaEspañola: This is the only one among the analysed dictionaries that does not provide any details about the period of time covered. The criteria used for the choice of neologisms are not clearly described, and neither are dates given and examples localized (Are they ad hoc creations?). This is an unsatisfactory practice which should definitely be banished from neological lexicography. Even a positive aspect, like the inclusion of verbal idioms, becomes problematic not only because of its treatment of idioms, but also because of the poor treatment of marked uses. For example, among the "expresiones" constructed with irse, irse a la mierda is marked as vulgar, but nothing is said of irse al carajo. Its dubious degree of neologicity is also problematic: irse al carajo, with the meanings this dictionary considers neological, is already documented by the CORDE corpus in 1961 (J. Mª Gironella), ir de culo in 1958 (M. Delibes), and irse a la mierda in 1938 (E. Serpa).
NDVocesUsoActual: The strongest point of this work, with 1032 neologisms per year, is also the weakest one -there is a proportional relationship between this figure and the fact that only one dictionary was used in the control corpus (the 2001 version of Real Academia Española). On top of this, and considering that the period under study spans over 12 years, and that it includes lexical items from Latin American Spanish (without any specific dictionary in the control corpus), it is not surprising to find such a high number of neologisms, as the probability of finding dubious cases of neologicity increases. Leaving this aside, this DN is finely structured and exceptionally well written, with a remarkable analysed corpus, abundance of examples and effective lexicographical definitions (including contextual information).
With criteria of selection of neologisms that are not clearly explained and 50 neologisms per year, the 24-year time span covered by NEspañolActual is too long for a dictionary that carries the adjective 'contemporary' in its title. The information of the articles is both concise and precise, and the high quality of definitions must be emphasized. Examples, only one per entry, have been carefully chosen. Variant forms, indicated in the entry, are alphabetically ordered and the example given can include either form (the examples given for aparcabicicletas or aparcabicis and coltán or coltan have the first form, whereas the examples given for anticaída or anticaídas have the second form). However, indications about lesser frequency of use of one of them (cf. §3.2) are exceptional, when in fact they should be systematically given.
DParaulesNoves: With a rigorous methodology, a comprehensive control lexicographical corpus and a time span within the limit of acceptability (8 years, with an average of 129 neologisms per year), this DN displays an excessively high presence of documented forms (28%), although this fact is appropriately indicated. Variant forms are correctly described and included as independent entries, with examples and specific indications. Examples are abundantly provided, and valuable complementary information is given, like typographic marks and, in general, indications when neologisms were coined by an author. Some definitions, however, need to be improved (cf. §4.1).
NPalabrasGalegas: With a similar methodology to the previous one, an 8-year time span and an average of 145 neologisms per year, its control corpus should include a dictionary published around the middle of this period (GDXL). The latter dictionary was in fact used, but only to mark those neologisms documented by it (a not inconsiderable 17%). The abundance of examples with the distinction of the two types of corpora used (newspapers and Internet), and especially the use of parallel textual corpora in order to record the first time the neologisms were documented, are practices that deserve praise.
NQuotidiani and ParoleNuove are two consecutive products of the same lexicographical project. With 1012 and 669 neologisms per year respectively, they constitute an excellent example of sampling contemporary Italian newspapers over short periods of time (five and three years respectively). The control Brought to you by | Universidade de Vigo Authenticated Download Date | 12/12/16 12:47 PM corpus of the first one only includes two dictionaries; the second one includes three large Italian dictionaries. This explains why NQuotidiani has 15% of documented neologisms (60% in other dictionaries of new words and 40% in general-purpose dictionaries) and ParoleNuove only 5% (a third of which can be found in dictionaries of new words). They systematically offer all sorts of information: abundant real examples including variants, authors are specified in ParoleNuove, information about the first recorded dates, and the components and formation procedure of all the neologisms are described in a schematic but rigorous fashion. This linguistic analysis is very valuable from the perspective of each new word individually and also from the global perspective, to understand the functioning of lexical renewal in the Italian language used in the media. Information on phonetics and details about thematic fields are missing, although the latter is partly compensated by the inclusion of details about newspaper sections. I would not hesitate to say that the authors of ParoleNuove have reached a level of excellence with this work that is very close to the ideal one would expect from a DN.

What Could An Ideal DN Be Like?
A DN should aspire to describe neologisms accurately, comprehensively and with the best possible organization so that searching is easy while at the same time offering abundant illustrations and explanations of their uses, as well as assessments of their neologicity and probable longevity.
What are, then, the main characteristics of a DN for me?

Choice of Terminology
Apart from the appropriate configuration a DN must have, it is essential that the neologicity of the chosen neologisms is guaranteed (cf. §2.1) and that users are in a position to assess them, with as much available information as possible, and without compromising on easy searchability. 1. Type of source: the corpus from which the neologisms were taken must be made explicit with absolute precision. It is important that corpora have a representative character, depending, of course, on the available resources and on the specific objectives one wants to achieve with a given DN. Written corpora are most often used for this purpose, but I would suggest an attempt should be made to introduce oral corpora, as long as some minimum conditions of standardization and quality can be guaranteed. In their contrastive study of neologisms taken from the radio and from the written press, Domènech, Estopà (2009) found interesting differences between them and suggested that a further, more exhaustive analysis should be obtained: "puede llevarnos a obtener conclusiones interesantes sobre los procesos de creación neológica" (2009: 63). 2. Time span: the period of time of the study would ideally be one or two years. I am aware that this condition would be hardly acceptable from the publishers' perspective (and not only from theirs), and besides, one must bear in mind that short periods of time are perfectly covered by online neological resources. In view of this, I suggest that the time span covered ranges from 3 to 5 years. Guaranteeing uniformity is of the essence, without putting their novelty in jeopardy too much. 3. Criteria for selection/exclusion of neologisms: this is the key starting point for any DN. I believe the point of departure should be a control lexicographical corpus, which should include the best, most comprehensive and recent dictionaries for each language, including also diatopic variation (cf. §2.2 and §3). Documented forms should not be included (or in any case, they can be introduced in a separate appendix), unless they have marks of neologicity in general-purpose dictionaries (in which case it should be appropriately indicated). New computer tools for extraction of neologisms have given us access to formal neologisms, but we must be especially vigilant for possible semantic neologisms through semantic change.

Conclusions
In this work, I have described three broad DN types: those developed by an author, by publishing houses and by Observatories of Neology. Even at the risk of oversimplifying, I can associate these three types with different objectives and methodologies. The first type is clearly influenced by the artistic-creative aspects of language; the second type favours expositional clarity and uses the standards of lexicographic processing of publishing houses, and the third type applies scientific and academic rigour to the lexical-semantical level of the new words in a language.
All of them, regardless of the languages in the study, were rigorously developed. Generally speaking, one of the strongest points of a DN is to explain the criteria used to identify a new form, and how filters are used to screen and profile the new words that are finally included as entries, so that the neological value of the terms (regardless of the initial criterion) is guaranteed (Crespo et al. 2008).
I have to highlight the remarkable quality of the lexicographic definitions and the documentation of orthotypographic variation of neologisms (of special interest for language policies, for example regarding loan words). The DNs in this study use lexical neology as their departure point to move beyond that and Brought to you by | Universidade de Vigo Authenticated Download Date | 12/12/16 12:47 PM focus on more complex units -and they do this satisfactorily -so much so that there are only some minor problems at sentence level.
Most DNs worked with a time span of less than a decade. This involves an important effort, as there is a need to work fast and relentlessly, keeping homogenous working procedures and being rigorous at the same time. Only this can guarantee that the DNs are up-to-date and that the information provided supports not only general dictionaries, but also the grammatical field of word formation.
I should also underline that some of these DNs include complete information about word-formation processes and indexes of word-forming elements. This is relevant as it helps overcome the potential limitations of using an in print format to study neologisms. These indexes of word-forming elements -and the information about word-formation processes -are a powerful tool to study new words in language in use in all their systematic complexity.
Indeed, these DNs show that for lexicography they are highly valuable tools and that, no doubt, in the 21 st century we can at last speak of a fully-fledged "neological lexicography".