The Concept of Lexicographic Condensation: A Review of and Perspectives on Digital Lexicography

: This study focuses on the concept of lexicographic textual condensation in order to determine its role in digital lexicography. Current interpretations of this concept, which was originally coined for printed dictionaries, are analysed. Special emphasis is placed on the relevant differentiation between the various textual levels in which the content of these works are distributed and on the difference between primary and secondary condensation. Examples from English – Spanish digital dictionaries will be explored in order to answer the question of how condensation is presented in these works and what types of components or items are particularly affected by it. Finally, meaning of condensation forms in current dictionaries based on the reading approaches applied to reference works is discussed. We support the argument that these forms are intrinsic to the type of text that we call a dictionary , and that the reasons behind condensation mechanisms are not limited to mere spatial constraints, but that these, together with other formal aspects, can make it easier to find the desired answers through a nonlinear reading process that is generally and legitimately applied to dictionaries.


The concept of lexicographic condensation
The framework of the general theory of the form of dictionaries (e.g.Wiegand 1984, 1996a, 1997, Wiegand and Fuentes Morán 2009) presents and develops the concept of lexicographic textual condensation (Wiegand 1996b, Bustos Plaza and Wiegand 2005-2006) as one of the main pillars for the formal description of its subject matter.This concept attempts to explain partially the particular, albeit not unique, form in which data are presented in standardized dictionaries.Indeed, one of the characteristics that identifies the lexicographic text 1 is the fact that, to a large extent, linguistic data are essentially expressed throughout the dictionary with forms of expression that are not part of natural language (Wiegand 2009).As a consequence, lexicographic condensation is described as the process that leads from a full text (a text showing complete cohesion and explicit syntax) to a condensed article text (with addressing as syntax, cf.Gouws 2015).This can be illustrated with a very simplified example.If we look for information about the word increase (which, in an attempt to simplify the question, will be discussed here exclusively as a verb) in a bilingual English-Spanish dictionary, we do not generally find entries in modern dictionaries with the following information: (1) increase is a verb (2) increase means aumentar but rather, e.g.
(a) increase verb aumentar or (b) increase v aumentar When comparing (1) and (2), which are presented in natural language, with (a) or (b), which present the characteristics of a lexicographic text, we can observe that (a) and (b) do not include "is a", that "verb" in (b) is abbreviated and presented as "v", and that "means […]" is omitted. 2 (a) and (b) are condensed sequences in non-natural language that try to provide the same propositional content than (1) and (2).Both (a) and (b) apply one of the most common condensing procedures: omitting; and abbreviating is also a technique applied in (b).Therefore, (b) is more condensed than (a).Together with omitting and abbreviating, other condensation mechanisms have been described: shortening, shifting, substituting, summarizing, and embedding. 3 However, condensation is described as a transformation process.If interpreted literally, this would imply that the text starts as a sequence in natural language (for example: "increase is a verb") which is then subjected to different transformations until it reaches the form in which the result is presented in the dictionary.However, we must admit that this principle can only ever be applied from a theoretical perspective (Rascón Caballero 2021: 104-106).That is, lexicographers in a real context do not generally formulate the contents in natural language and then transform them through condensation procedures in order to include the data in the dictionary according to the required standard format.Therefore, an actual condensation process is not taking place.However, the purpose behind the application of this concept goes beyond this principle.The objective is to determine formally the propositional content of the sequences that are transformed -in other words, the data that need to be transmitted -and to contribute to the description, development, or assessment, among others, of the corresponding forms of representation of these data in the dictionary.This, in turn, provides a tool that can be applied, for example, to perception studiesin other words, to studies on how the transmitted data are interpreted and on what potential sources of error can be found in these interpretations.Therefore, the strengths in the application of this concept include its contribution to the detailed description of the differences between natural language and the forms of expression used in the dictionary and, with it, to an accurate description of the different levels of lexicographic textualization (e.g.Wiegand 1996a).In sum, this type of description makes it possible to establish the initial guidelines of a partial methodology for the planning, design, comparison, and assessment of dictionaries on a formal plane -which is intrinsically connected to their contents (Wiegand 1998).This provides a base to study the way in which users infer the information that is conveyed through the dictionary, as supported by Nielsen (2002: 597): The lexicographical information can only be inferred from the article when the user successfully establishes the relation between two items, e.g. the item giving the meaning paraphrase and the item giving the form of the lemma sign.Therefore, the way in which the items have been written is of the utmost importance in establishing this relation, and this may be analysed in terms of textual condensation, which also has a direct bearing on the readability of an article.
The incidence of condensed components in a lexicographic article, as can be seen in (a) and (b) allows us to classify it as being more or less condensed, as can be also seen in the following two examples: http://lexikos.journals.ac.za; https://doi.org/10.5788/33-1-1824(Article) The example in Figure 1 presents "intransitive verb" without an abbreviation (and therefore, without condensation), whereas the example in Figure 2 shows "V.intr.", which is an equivalent propositional content to the previous one, but which is presented here in its abbreviated (condensed) form.In any case, it is necessary to consider that all the dictionary articles, as they are currently known and according to the basic conventions of the object that we call a dictionary, always present a certain degree of condensation.Even if all the components in the lexicographic article were presented in their uncondensed formwhich is not at all common -we always find at least some omissions: the components are not linked to each other through the procedures of natural languages, as can be seen in Figures 1 and 2. There are no sequences such as "an equivalent for increase in Spanish is …", "the meaning of increase is …", "increase is an intransitive verb …" or "increase, as an intransitive verb, means …".No, as observed, the corresponding data are formalized through different procedures that are typical of dictionaries and that allow us to identify them as such.
Moreover, the process that we call lemmatization is a type of condensation that is typical of lexicography and characterizes it.For example, in all the dictionaries in which the word is included, increase is a lemma sign that, through a process of selection -and, consequently, the omission of all other potential elements -represents the entire paradigm of the verb. 4 Therefore, it is essential to establish a category for a type of condensation that will be described as primary and that encompasses (1) the procedures that determine the form of the lemma sign, and (2) the omissions of expressions in natural language that join the different components of the article, as seen before (expressions such as "the meaning of … is …" or "… is an intransitive verb …").
These omissions can also be seen in instances of non-lemmatic addressing.For example, bilingual dictionaries may show the following: (a2) increase verb aumentar verbo or (b2) increase v aumentar v Both "aumentar verbo" and "aumentar v" show the omission -a characteristic of modern dictionaries, as has been pointed out -of the element of natural language that would join syntactically "aumentar" with "verbo" (a2) or with "v" (b2).
Therefore, this is what we call a primary type of condensation that has been conducted through omission.
The other varied forms of lexicographic condensation which are generally decided when planning the design of each specific dictionary and which may be motivated by the linguistic conventions of each language -such as the abbreviation systems -may be referred to as secondary condensation.
Currently, primary condensation identifies the modern lexicographic text as such.Secondary condensation is just a possibility.It seems that those who insist that condensation is no longer necessary in digital dictionaries refer to this secondary form (Wolski 2020). 5 The result of the condensation does not necessarily appear through linguistic signs, but also, for example, through icons and symbols, such as asterisks, arrows, etc., or through a combination of linguistic and non-linguistic signs.That is the case, for example, in the printed dictionary Oxford Study, targeted at Spanish-speaking English students: Here, the symbol of an arrow with the letters "LOC" inside it is revealed as "locuciones y expresiones" [phrases and expressions], while the same symbol with the letters "PHR V" inside means "sección de phrasal verbs" [phrasal verb section].The symbol of a key inside a circle means "información sobre uso de palabras de uso más frecuente" [information on the usage of the most frequent words].In the example in Figure 2 there are symbols that represent the flags of the United States and the United Kingdom, which correspond to "American English" and "British English", respectively, as can be seen in Figure 4 by hovering the cursor over the symbols.4, the explanations "American English" and "British English" are visible by hovering the cursor over the flags, as previously shown.The flag symbols are, therefore, items that are presented in a condensed form, and which become uncondensed through a textual level that is accessed by hovering the cursor over them.Another example of un-condensation that is visible by hovering the cursor over a condensed form in the initial level can be found in Wordreference E-S: In Linguee E-S (in which the textual level is not developed by hovering), the information regarding pronunciation (a recording) is accessed by clicking on the loudspeaker symbol.This is a way to expand information through a reference: the initial level, through the symbol of the loudspeaker, refers to a different textual level in which, through another reference, the user can access the oral reproduction of the word.
As we can see, changes in the textual level are particularly used to expand the information (references) or to un-condense some of the elements that are initially condensed.The example in Figure 7 is a very common type of reference in digital dictionaries.
Although it is generally possible to establish a clear difference between the expansion of information achieved through references and un-condensation, we can also find some less precise cases in which both concepts are entwined.The following example can be found in Cambridge E-S: , where we can find the full list of these elements and, in several cases, a short explanation of their meaning (Figure 9).This is, therefore, an un-condensation procedure that presents the full form of items.In addition, the expanded information included in the original lexicographic article can be found in that same location.
[…] Trends in digital lexicography (English-Spanish) This section discusses the most characteristic types of items that include condensation and the type of procedures they follow in the selected dictionaries.Indeed, in some of these dictionaries we can find, apart from primary condensation forms, instances of secondary condensation, particularly in (1) morphosyntactic items, such as v (verb) or pl.(plural) -probably the most classical and visible type -and (2) the way in which disambiguators are presented, and occasionally the definitions, equivalents and examples.The pertinence of including these items or the forms in which they are presented are not discussed here, but the fact is that we can find a range of different decisions in the examples from the selected dictionaries.An overview of the most significant cases will be presented and the way in which condensation has been applied in the articles corresponding to the verb increase will be analysed.

Morphosyntactic items
Included here are all the items that refer to different classes of words, syntax (such as transitive/intransitive), and morphological items (e.g.conjugation models).In Collins E-S, Cambridge G. E-S, and Longman E-S, we can find information presented in a non-condensed form regarding the fact that increase is a verb, and also regarding its transitive or intransitive nature.Figure 1 showed the example from Collins E-S ("intransitive verb").Figure 10 shows two more examples.The dictionary Nglish E-S does not condense the word verb, but in this case there is no information on the transitivity or intransitivity of the word (Figure 11).Examples where the form verbo or verb is not condensed but the information on its transitivity or intransitivity can be found in a condensed form in Cambridge E-S (Figure 8) and in Linguee E-S (Figure 12).The information on the transitivity or intransitivity of the verb is indicated through the condensed form "(sth.)".In this case, the mention to the object is abbreviated (sth., as an abbreviation of something), and the parentheses indicate that this is an optional element.The item is not linked here to another textual level and, therefore, it is a condensed element that does not give access to its explanation or un-condensation.
Both Wordreference E-S and Pons/Oxford E-S include condensed items on the category of the verb and its transitivity or intransitivity.Figure 13 shows an example from Wordreference E-S, which is discussed below, together with its conjugation forms (Figures 17-20).In Pons/Oxford E-S, hovering the cursor over the condensed part of speech shows the following pop-up information (Figure 14): The forms of inclusion on the verb conjugation also show that different decisions have been made.Pons/Oxford E-S also presents a hyperlink to a conjugation table that provides extensive information related to the article through a process similar to referencing (Figure 15).In Linguee E-S (Figures 7 and 12), next to the word "verbo" can be found "(increased, increased)", which, as users may deduce, are the forms of the past simple and the participle (in no specific order).By clicking on "increased" more information is available (Figure 16): The classical verb conjugation forms (in this case increased, has increased, is increasing and increases, in that order) can also be found in Nglish E-S (Figure 11), although without any hyperlink, explanation or additional information.
The case of the tool Wordreference E-S is unique and will be discussed separately.The items on the word category and conjugation are condensed and uncondensed in the same textual level.This example presents the uncondensed items noun and verb in the first line, and the corresponding condensed items (v. and n.) in the second and third lines. 6These are, therefore, items with the same propositional content and presented at the same level, but in two different forms.The abbreviated items v. and n. are not explained by hovering the cursor over them or by clicking.The classical forms from the conjugation model are included (increases, increasing, increased and increased), as in other dictionaries, but in this case, they add an indication of the form (or person) for each of them.Once again, "person singular" is not condensed, but the forms "pres p", "past" and "past p" are partially condensed.In this case, condensed forms are presented in full in a second textual level that can be accessed by hovering the cursor over them: Here, apart from the full form (1), we can find an explanation (2) of the meaning of "present participle", "past simple" and "past participle", respectively, and examples (3) for each of them.These are different types of information, with two different forms of addressing, in the same pop-up.In summary, as we can see, there is a wide range of decisions that have been taken regarding the presentation of morphosyntactic items.

Disambiguators and other condensed components
Semantic annotations act as disambiguators and establish differences between the meanings and equivalent alternatives.In Collins E-S (Figure 21), the annotation [number, size, speed, pain] is partially condensed through the omission of the characteristic linking elements from natural language, and through a linking procedure with commas; and it may be interpreted as a simple group of generic items of a cotext.

Figure 21: Disambiguators in Collins E-S
The same procedure can be seen in Pons/Oxford E-S (Figure 2) and in Longman E-S.
In this last case, this does not appear in the English-Spanish section, but in the Spanish-English one (Figure 22), which may be due to the didactic approach of this dictionary.As is known, some bilingual dictionaries also contain definitions. 7The following example from Cambridge E-S (Figure 23) includes a case of condensation through embedding.This example shows an embedding mechanism which is not uncommon in definitions, but which is less frequent within examples.The sentence "to (make something) become larger in amount or size" embeds two definitions into one: "to become larger in amount or size" and "to make something become larger in amount or size".A similar procedure can be found in the second example: "The cost of the project has increased dramatically/significantly since it began".As we know, the slash replaces "or" in natural language.The same type of formalization can be found in Collins E-S (Figure 21: "to increase in weight/volume/ size/value …") or Cambridge G. E-S (Figure 24).This example shows that the definition is not included in this procedure, because it is presented in natural language at all times, using "or".We can also find slashes in the presentation of translations of examples into Spanish, as can be seen in Longman E-S (Figure 25) Finally, we must mention another significant factor: the presence of identical condensed items ("US", second and third lines), with a difference that can only be observed by hovering the cursor over them:

Condensation in digital dictionaries?
The presence of condensed forms in the printed format was markedly determined by the need to save space and provide as much information as possible, and it was considered necessary in order to optimize the printing surface.This may be a plausible explanation for the use of abbreviations and acronyms, as well as other condensation forms such as omission or embedding.The lack of space is a reasonable explanation for the fact that data in printed dictionaries are presented in formats that diverge from natural language.
Over the last decades of the 19th century and the first decades of the 20th century, many social and commercial changes have expanded the scope of interlinguistic communication, and there was a growing need for dictionaries that were easier to transport and use.These changes paved the way for the arrival of more reduced formats that joined large dictionaries, particularly those known as "pocket dictionaries". 8This decrease in size did involve including fewer entries and less information, but also led to the systematic development of different methods to save space and, as a consequence, of the forms that we now know as condensation.Dictionaries evolved to become increasingly standardized texts, and only rarely can we find, in the last decade of the 20th century, some examples of a slight change of trend in pedagogical lexicography.
However, in modern printed dictionaries, the question of space is combined with other aspects related to the forms in which the book is consulted, the ways in which it is read.In fact, textual density in dictionaries is not only motivated by the need to save space -which is more evident in printed dictionaries -, but also, as many other current means of communication, by the search for greater dynamism and effectiveness in the transmission of the message (Torres del Rey and Fuentes Morán 2013).In fact a dictionary is not a text meant to be read from left to right and from top to bottom (or in the direction required by each language).The same applies to each of the lexicographic articles it contains.This does not occur now, in digital dictionaries, just as it did not occur earlier, in printed formats.
Let us remember, in line with Rowe (2013), the difference between at least two reading modes ─ linear and tabular reading: 9 The act of reading has numerous and highly distinct functions and attitudes […].Two basic types of engagement with a text may be identified, however, which are generally termed linear and tabular reading.Linear or intensive reading characterises the way we consume narrative fiction, [...] the reader of such works is typically highly absorbed in their storylines, borne along almost automatically by the temporal and causal narrative connections between sentences, paragraphs, and chapters.The tabular mode of reading, however, does not correspond to the storytelling form […]. In this case, reading is interrogative, seeking information about a specific subject, and this is reflected in the formats of tabular textsencyclopedias, dictionaries and other reference books.Such works employ a number of organisational strategies -including alphabetisation, block spacing of text, section headings, and so on […], all of which are designed to facilitate the readers' search for specific information as well as to direct them towards the text's sources or related material if necessary.
In fact, the typical reading mode when consulting a dictionary is tabular, according to this classification, but it is also referred to as "scanning" (Rosenwald 2014) or hyper reading (Hayles 2010), for example, and its true purpose is to locate and obtain specific information or, in fewer cases, more general information about a lexical unit.We navigate across the dictionary and our eyes scan the lexicographic article until they have found (or they believe they have found) the information they were looking for.The process may continue if the previous results were not satisfactory, or it may end at that point, regardless of whether the article contained more information than what was actually assimilated by the reader.In dictionaries, as in many other texts, information is structured with more or less advanced formal characteristics that enable this type of reading.These characteristics are position, size, font, style, color, and contents.With contents were refer to the components of the lexicographic article as they are presented in the dictionary, but regardless of their position, size, font, style, or color (because these elements, as structural indicators, are not meant to convey linguistic information and propositional contents, but to support and clarify the structural organization of the lexicographic article).
Finally, as stated by Krug (2014) when he tried to answer the question "Why do we scan?": [1] We're usually on a mission.Most Web use involves trying to get something done, and usually done quickly.As a result, Web users tend to act like sharks: They have to keep moving, or they'll die.We just don't have the time to read any more than necessary.
[2] We we don't need to read everything.On most pages, we're really only interested in a fraction of what's on the page.We're just looking for the bits that match our interests or the task at hand, and the rest of it is irrelevant.Scanning is how we find the relevant bits.
[3] We're good at it.It's a basic skill: When you learn to read, you also learn to scan.We've been scanning newspapers, magazines, and books -or if you're under 25, probably reddit, Tumblr, or Facebook -all our lives to find the parts we're interested in, and we know that it works.

Conclusions and perspectives
The practices that were originally justified in printed formats because of the spatial constraints acquire a new dimension and different characteristics in digital lexicography.The revision of the essential aspects of the concept of lexicographic condensation allows us to reconsider its relevance and to qualify and modify it in the field of digital lexicography.This is not due to the inertia of classical procedures, but to the need to adapt the works to the changing characteristics of the surface -which can now be visualized in all sorts of screens, from the smallest mobile phone to the largest monitors -and to the different textual levels that can be defined.However, beyond these considerations, current formats promote or must promote and enable reading modes that were already characteristic of dictionaries and are now inherent to digital tools as a whole.
We refer here to dictionaries and lexicographic articles as text, as usual, in a generic sense.We are aware, however, that in a strict linguistic sense, these works do not completely meet the requisites for textuality (cf. De Beaugrande andDressler 1981, Bernárdez 1982).2.
We will not discuss other aspects that are intrinsically related to the entries, such as the characteristics of the format (bold, italic) that act as a support for information and, therefore, for legibility and readability (see Sections 3 and 4).

3.
A summarized and exemplified account of lexicographic condensation can be found in Wolski 1989, on monolingual dictionaries;and in Wolski 1991, on bilingual dictionaries.The form in which each of these condensation types and methods can be described may be consulted in Wiegand 1996a, 1996band Wiegand 1998, particularly.The works that analyze the condensation procedures of specific dictionaries include Nielsen 2002, Gouws 2005, and Fuentes Morán and Pradas Macías 2009. 4.This work will not establish a difference between inner textual condensation -which concerns all those lexicographic partial texts containing a carrier of the guiding element, in particular dictionary articles and index entries -and outer textual condensation -pertaining to the carriers of the guiding element such as lemmata and sub-lemmata, for example.This difference is particularly relevant for printed dictionaries.In addition, a qualitative approach, rather than quantitative, will be applied here, therefore, the different degrees of condensation are not described here.5.
All the dictionaries used here to provide examples show instances of primary condensation.As we shall see, they also present secondary condensation to different extents.6.
Other possible incoherences regarding position and structure can also be seen, but they will not be discussed here.7.
This study will not discuss whether these dictionaries are strictly bilingual, bilingualized, semibilingual, etc.They will be treated as bilingual dictionaries because they contain two languages which are analyzed, and our teaching experience indicates that students do treat them as bilingual dictionaries.8.
Observing and interpreting the evolution of dictionary formats through history can only be done with a broad perspective.This evolution is influenced by a wide range of factors, among which commercial, ideological and socio-cultural ones should not be overlooked.Some interesting notes on this can be found e.g. in Cowie 1998and 1999, Hartmann 2000or Collison 1982.9.
There are several works that cast light on this topic and present different advances.We particularly recommend Durant 2017 because of its analytical literature review on reading forms in the digital era within the framework of the widely known Charleston Briefings.

Figure 1 :Figure 2 :
Figure 1: Partial article of increase from Collins E-S

Figure 3 :
Figure 3: Fragment of the instructions in Oxford Study

Figure 4 :
Figure 4: Pons/Oxford E-S explanation of the flag symbols In contrast with the limitations of the two-dimensional surface of printed dictionaries, digital dictionaries may present up to three characteristic textual levels: (a) An initial level (extended by means of a more-less folding system) (b) A second level that is accessed by hovering the cursor over a component in an article (c) A third level that is accessed by clicking a component in an article Let us briefly discuss some examples of these three levels.(a)The initial level is sometimes expanded or collapsed through the more-less feature (or a similar procedure), as can be seen in Wordreference E-S:

Figure 5 :
Figure 5: Partial article of increase in Wordreference E-S (b) In the example in Figure4, the explanations "American English" and "British English" are visible by hovering the cursor over the flags, as previously shown.The flag symbols are, therefore, items that are presented in a condensed form, and which become uncondensed through a textual level that is accessed by hovering the cursor over them.Another example of un-condensation that is visible by hovering the cursor over a condensed form in the initial level can be found in Wordreference E-S:

Figure 6 :Figure 7 :
Figure 6: Partial article of increase in Wordreference E-S with information that is presented in a non-condensed form for UK through hovering

Figure 8 :
Figure 8: Partial article of increase in Cambridge E-S By clicking on [I or T] -hovering the cursor does not have any effect -we can access the section "Labels & Codes", where we can find the full list of these elements and, in several cases, a short explanation of their meaning (Figure9).This is, therefore, an un-condensation procedure that presents the full form of items.In addition, the expanded information included in the original lexicographic article can be found in that same location.

Figure 9 :
Figure 9: Partial article of increase in Cambridge E-S

Figure 10 :
Figure 10: Partial articles of increase from Cambridge G. E-S and Longman E-S

Figure 11 :
Figure 11: Partial article of increase from Nglish E-S

Figure 12 :
Figure 12: Partial article of increase from Linguee E-S

Figure 13 :
Figure 13: Partial article of increase from Wordreference E-S

Figure 14 :
Figure 14: Partial article of increase from Pons/Oxford E-S with an explanation for "V." and "intr.",respectively Although in Pons/Oxford E-S the item "intr."Always appears next to "V.", "intr." is uncondensed as "verbo intransitivo" [intransitive verb].The forms of inclusion on the verb conjugation also show that different decisions have been made.Pons/Oxford E-S also presents a hyperlink to a conjugation table that provides extensive information related to the article through a process similar to referencing (Figure15).

Figure 15 :
Figure 15: Fragment from Pons/Oxford E-S with part of a conjugation table

Figure 16 :
Figure 16: Partial article of increased from Linguee E-S

Figure 17 :
Figure 17: Partial article of increase from Wordreference E-S

Figure 18 :
Figure 18: Partial article of increase from Wordreference E-SThe section "Inflections of 'increase'" also shows different items, related to inflection, and presented in both condensed an uncondensed form in the same textual level:

Figure 19 :
Figure 19: Partial article of increase from Wordreference E-S

Figure 20 :
Figure 20: Partial article of increase from Wordreference E-S

Figure 23 :
Figure 23: Embedding in a definition in Cambridge E-S

Figure 24 :
Figure 24: Embedding in a definition in Cambridge G. E-S

Figure 25 :
Figure 25: Translation with slashes in Longman E-S

Figure 26 :
Figure 26: Partial article of increase from Wordreference E-S