ABSTRACT: A dictionary of the language of food, as a dictionary of a special sublanguage in the Serbian language, does not exist to this day. 
Too ordinary or too prone to changes, the description of the lexicon of this sublanguage escapes the attention of language researchers, and confuses lexicographers. 
The paper analyzes examples of lexicographic processing of individual lexemes of this language and shows that the description of the culinary lexicon cannot be solved by standard lexicographic procedures. 
A digital corpus consisting of different types of texts describing the language of food is presented, including cook-books, ethnographic and historical studies, literary works, etc. 
It is shown that the language of food evolves rapidly and continuously, and its evolution is closely related to changes in society. 
Finally, the author advocates the creation of an encyclopaedic dictionary of the language of food and indicates its basic properties. 
KEYWORDS: language of food, domain corpus, lexicography. 
The language used when speaking about food is so quotidian, so ordinary that almost everything that is said becomes immediately understandable to everyone. 
Such a view relies on the fact that a plate contains, except in extraordinary circumstances, the usual dish, whose ingredients are well-known, forming part of the “traditional” cuisine. 
It is likely that these are the reasons why the investigation of the language of food and especially its temporal, spatial, cultural and social evolution has been largely overlooked in our midst. 
This language, or more precisely sublanguage in Harris’ sense (Kittredge and Lehrberger 1982), at first glance features a limited lexicon and a narrow choice of syntactic constructions, which intensifies the impression that its nature is simple. 
However, unlike numerous other sublanguages, the language of food is constantly changing, above all, under new cultural influences and changes in food fashions. 
On the other hand, despite all of its simplicity, the transfer of culinary content from one language to another is exceptionally complex, since the conceptualization of food differs from language to language, not only in the structure of the dishes that get prepared, but also in what is considered to be acceptable (and customary) choice of ingredients and dishes. 
This transfer also involves a constant process of naming the foodstuffs and dishes adopted from other places (Радоjичић 2015). 
In order for something to become food, it must have (acquire) its linguistic expression. 
Each element constituting this language, starting from the ingredients, moving to the ways of their preparation and ending in the prepared dish has its own name, thus naming and differentiating items in the language of food in a way precedes the food itself. 
Let us note immediately that the lexical elements of this language are, as a rule, abstract notions in the sense that they are devoid of their features. 
As we will see, lexicography is faced with this nature of the lexis of food when a culinary term needs to be defined. 
For example, the lexeme tomato, when it appears in a recipe, denotes just a general concept and in most cases does not provide precise information as to the kind of tomato. 
However, this fruit can have additional attributes which will yield different results when preparing food and that aspect evades lexicographers’ attention. 
A lexicographic description will not, except in extraordinary circumstances specify the kind of tomato – jabučar (apple-like tomato), volovsko srce (beef tomato), čeri (cherry tomato), etc. – or the meaning of the expression tubed tomato paste, peeled tomatoes, tomato soup, tomato broth, etc. 
A description of the relation between the meaning of these words and their usage (at any given moment) relies on the collective notions related to the meaning of the concept tomato. 
In this paper, we will be interested in the issue of evolution of the language of food in Serbian ever since the appearance of the first cookbooks to the present day. 
By defining the issue in question in such a way, the description of the language of food gets narrowed down, primarily, to townspeople’s cuisine in the regions populated by Serbs, as illustrated by cooking handbooks. 
Beyond that framework, countryside eating habits that are believed to follow tradition and customs, in terms of food preparation, experience changes in an indirect and belated manner. These changes, coming from the urban cuisine have to do both with the technology of preparing food and the range of foodstuffs and dishes. 
This claim can be supported by the depictions of the contemporary eating habits in rural areas (Радуловачки 1996; Лазић 2000; Радовановић 2011) that attest to the integration of foodstuffs, dishes and cooking techniques present in antique cooking handbooks that older ethnographic sources (Троjановић 1983; Ердељановић 1908; Ђорђевић 1928) were not familiar with, though. 
Cookbooks in Serbian, often featuring words such as “Serbian” or “national” in their titles resulted, above all, from the attempts to express in that language culinary notions coming from the areas with developed cooking skills and traditions. 
Thus, Midžina’s handbook Поповић-Миџина (1891) named Veliki srpski kuvar (Big Serbian Cookbook), provides primarily a description of the Viennese cuisine, while Veliki narodni kuvar (Big National Cookbook) by Aleksandra Рустановић (1938) endeavours to evoke dishes belonging to the French and Russian cuisine in the numerous recipes in Serbian. 
It is clear that texts of this kind are faced with the problem of naming new and unfamiliar concepts referring either to ingredients or prepared meals. 
In that sense, the description of the language of food bears witness to the process of modernization and internationalization of the domestic cuisine. 
By comparing the content of cooking handbooks with the range of fast or convenience food on offer in Belgrade in early 21st century (Витас 2018) we can see that the contemporary range of “fast food” dishes, which are nowadays viewed as being ordinary, sometimes even traditional were unfamiliar in the first half of the 20th century and later, as well. 
Such nature of the language of food was recognized as early as in Big Serbian Cookbook by Katarina Popović Midžina, who was quick to make the following point in the introduction to the first edition: 
Let’s say a few words now about “artificial expressions” too, namely, foreign words that we must use out of necessity in this book of ours. 
Since, we Serbs inherited very few dishes and desserts, no one will hold it against us or find fault with us if we use German, French and foreign names in general. However, we will also keep those names that were passed on by our ancestors and remained unscathed in the process. 
Today’s generation is accustomed to affluence and luxury and keep it up in terms of dishes too. 
There are thousands of new dishes for which we have no names, despite the fact that we have come pretty long way as far as books are concerned. 
A good and comprehensive Serbian cooking dictionary that we could consult as the need arises is still missing (Поповић-Миџина 1891, 11). 
The “artificial expressions” mentioned here refer to the neologisms coined by Midžina in order to render into Serbian primarily recipes belonging to the Viennese cuisine. 
The text of this book itself provides an indication as to what could be featured in a “good and comprehensive Serbian cooking dictionary”. 
Striving to describe her recipes, Midžina resorts to giving synonyms, or foreign equivalents, so as to explain to the users what the recipes are about, like in the examples below:  
– Podaj na trpezu sa zejtinom - uljem i sircetom´ - octom (Put it on the table with oil - oil and vinager - vinager) (44); 
– Bela bundeva ili koraba - келераба (White squash or turnip cabbage - kohlrabi) (360); 
– Arapska smola - gumiarabikum (Arabian resin - gum arabic (357); 
– O ukrasima za govedinu. Rindfleisch – Garnirung (On Beef garnish. Rindfleisch – Garnirung) (73). 
The phenomenon of the evolution of the language of cooking is discussed in the 1911 fourth edition of this Cookbook (Поповић-Миџина 1911, 367), as well, where the publisher points out: 
The needs of a contemporary household require that it too should be recast in a more modern guise, and in order to do it, we have purchased permanent publishing rights to this book from the legal heirs of the late Katarina Popović-Midžina, so as to be able to edit this useful local book independently, providing it with a direction followed by the forward-looking housewives today. 
Less than 20 years after the appearance of the second edition, at the beginning of the 20th century, the new publisher thoroughly changed and modernized the content of this book. 
Therefore, even at the time when “culinary fashions” were not placed under pressure of globalization and the internet, cooking and the related culinary products had been subjected to rapid changes. 
An insight into the scope of the changes can be obtained by comparing the 3rd and 4th editions (Витас 2022), namely, there are new recipes for the dishes having the same name, some dishes are “outdated” and are no longer featured in the new edition, while there are also some new hereto unknown (or undescribed) ones. 
A century and a half after the first edition of Midžina’s Cookbook, the language of food still awaits its own “good and comprehensive” dictionary. 
However, a dictionary of this kind must not be a mere listing of the vocabulary used in the domain of cooking. Instead, it should also provide a description of meaning that in addition to the contemporary and “usual” meaning includes its evolution from the moment of appearance of the first cooking handbooks. 
The descriptions of the lexis of cooking in the local lexicographic practice and its use in the wider context of talking about food i (Section 2) will be the first to be examined in this light. 
In order to make these different descriptions of the language of food the topic of an analysis relying on the contemporary information technology methods, we will provide a brief outline of the corpus consisting of Serbian cooking hand-books published from the mid-19th century to the present, as well as other texts depicting our eating habits and highlight the issues related to corpus organization and exploitation (Section 3). 
We will wrap up (Section 4) by providing a framework for building an (encyclopaedic) dictionary of cooking and outline the directions of further research (Section 5). 
The lexis of the language of food has been described primarily in the large dictionaries of Serbian (RSANU 1959–2019; RMS 1976; RSJ 2007). 
These dictionaries provide, for the most part, depending on the time of publishing, a description of an inventory of Serbian entries playing part in expressing cooking-related content. 
But the impact of that quiet evolution of the language of food results in the lexemes described in them not featuring new concepts and new meanings. 
The qualifierkuv(anje, kulinarstvo) (cooking, culinary art) has been proposed for the purpose of describing the lexis of the language of cooking, but as it has been shown in (Стиjовић, Станковић, and Сабо 2017; Радовановић 2017), assigning this qualifier in general dictionaries creates a number of problems and dilemmas. 
A search for the entries marked with this qualifier yields only a small number of them, e.g. aspik (aspic), mari-nada (marinade), medujelo (relev´), grenadirmarš (Grenadier March), krompirača (potato pie)..., mostly originating from foreign languages or regional naming of dishes. 
The description of the entries that are indispensable, a widely known fixture of the language of cooking does not include this qualifier. 
For example, the entrygovedina (beef in (RMS 1976) is described as beef meat without the qualifierkuv, while the entry rinflajš (beef garnish) in the same dictionary is jelo od govedeg mesa, kuvana govedina (beef dish, cooked beef ) having the qualifier kuv. 
Similarly, the entry musaka (moussaka) does not feature the qualifier kuv in (RSANU 1959– 2019), despite being described as a dish made of chopped meat, potatoes or other vegetables and eggs browned in the oven. 
Separating part of the wealth of lexical information in Serbian defining the concepts relevant to the language of cooking is possible by using other qualifiers, such as agr(iculture), bot(any), etc. (Стиjовић, Станковић, and Сабо 2017). 
This heuristic procedure, however, yields an approximate result only, as shown by the following example. 
Fruit is described as a botanical notion, while different varieties of a fruit are classified as an agronomic one in (RSANU 1959–2019). 
So, the entry jabuka (apple), namely, its basic meaning gets described in the same source as a botanical term: 
jabuka ž 1. bot. voćka Malus iz f. Rosaceae sa mnogo vrsta, koja rada mesnate, okruglaste za jelo ukusne plodove; plod te biljke. (apple f. 1. bot. fruit from the Mal us family with many varieties having succulent, roundish, edible and tasty fruits; fruit of that plant.) 
However, an apple variety – kožara (lit. leathery (skinned)) – gets described as an agronomic term: 
kožara ž 2. agr. a. vrsta jabuke sa debelom korom. (kožara f. 2. agr. a. an apple variety with thick skin.) 
Note that this definition of kožara is ambiguous: if in the definition of kožara, apple gets replaced by the above definition of apple, thick skin can refer both to the apple tree and its fruit. 
On the other hand, the qualifiers agr, bot and the like are also used when describing entries not belonging to the language of food, since their referents are not consumed as food, representing agronomic terms for certain plant growing procedures or plant illnesses, for example. 
The only excerpted culinary source in the (RSANU 1959–2019) material is the third edition of Midžina’s Cookbook. 
The examples from this handbook illustrate more than three hundred entries, but only a couple of them like aspik (aspic), bajcovati (to soak), madžun (grape syrup), mutvara (schmarrn), nabockati (to prick), nazadevati (to stuff ), zavariti (to marinade), zavezača (ravioli-like dish) feature the qualifier kuv. 
Entries like batak (drumstick), belanjak (egg white), bubrežnjak (loin), devdir (colander), garnirati (garnish), ispržiti (fry), jetrenica (beef steak), etc. do not include the qualifier kuv, although their usage is illustrated by an example from Midžina and it is clear that they belong to the language of food (too). 
A particularly interesting example is provided by the entry jetrenica, a neologism introduced by Midžina so as to replace German lungenbraten or beef steak (Витас 2022). 
This entry is defined and illustrated in RSANU (1959–2019) in the following way: 
jetrenica: 2. v. bubrežnjak (1 a) - Najslade je jetrenica - lungebraten. (jetrenica: 2. see bubrežnjak (1a) - jetrenica - lungebraten is the tastiest.) 
However, bubrežnjak (1a) that in Midžina refers to a different piece of meat (with different German equivalents) is defined in RSANU (1959–2019) as kidneys with tallow and meat around them and illustrated by an example from Midžina: Uzmi od mladog govečeta bubrežnjak (Take the loin of a veal).  
A different example is provided by the entry mesnjača whose principal meaning is defined in RSANU (1959–2019) as: 
mesnjača ž pokr. 1. vrsta kobasice od mesa; isp. mesnatica, mešnjača, (mesnjača m. regional. 1. a kind of meat sausage; cf. mesnatica, mešnjača, / meaty sausage,) 
which is not found in Midžina or the majority of other books on cooking. 
For instance, in (RK 1915; Мирковић 1922; Марковић 1959), the expression mesn(at)a kobasica (meat(y) sausage) is used for this kind of sausage. 
This expression appears in RSANU (1959–2019) not as a separate term, but as part of examples for the entries nadenuti (to lard) and krtinast (meaty – without bones) and it seems most likely to be an unrecorded synonym of mesnjača. 
But, in the cookbook (HrL 1878), whose author is Midžina’s contemporary, and in (Рустановић 1938) as well, mesnjača is the only term found. 
Automatic recognition of the entry mesnjača as belonging to the domain of food, based on the above-mentioned definition and having the same meaning as mesna kobasica seems like a time-consuming, if not an impossible task, to say the least. 
Some entries appearing in the third edition of Midžina’s Cookbook, such as butertajg (butter puff pastry), demermaviš (dough cutter), katarceltl (caramel candy)..., are among the ones omitted from the culinary terms in the above dictionaries. 
In the fourth thoroughly revised edition of that cookbook (Витас 2022), some of these words are featured, including the illustrative examples that define them: 
– ... onaj ureckani mali točak, kojim se seče testo („demer-maviš“) (...that jagged little wheel used for cutting dough “demer-maviš“) (Поповић-Миџина 1911, 498);  
– ... umešena već gotova testa od slatkog masla („butertajga“, Butterteig) (...premade sweet butter dough –Butterteig) (169). 
On the other hand, katarcelt was illustrated as early as in the third edition, in a separate chapter, titled Bonboni-katarceltl (Candy-Caramel).  
Excerpting the fourth edition (Поповић-Миџина 1911) means finding words that enriched the lexicon of cooking in the meantime, namely, the words Midžina herself, had not been using in previous editions. 
The entries bardirati, blanširati, frikasirati, haše, dobatnjak, karičice, raznolije, absent from (Поповић-Миџина 1891) are among the examples of this kind. However, they are used in the fourth edition. The definitions or equivalents in foreign languages are stated in many cases: 
– svaki njen komad pre zgotovljavanja beliti („blanširati“ blanchiren) (each piece is to be blanched before cooking (“to blanch”, blanchiren)); 
– Bardirati (Bardiren) znači obložiti meso slaninom (to bard (bardiren) means to cover meat with bacon); 
– Haše je jelo od iseckana mesa ili ribe (Hach´e is a dish made of meat or fish cut into pieces); 
– Frikasirati znači omanje komade mesa preliti začinjenim umokcem koji se gotovi od kuvana iseckana kroz sito procedena mesa (fricasser means to soak small pieces of meat in a dressing with spices made of cooked and chopped meat drained in a colander); 
– Buhavice („krapfne“) sa cimetom (doughnuts with cinnamon);  
– jedan batak sa karabatkom ili dobatnjakom (a drumstick with the upper leg or thigh of poultry);  
– разнолиjе (Mixed Pickles); 
– ispržene karićice luka ili karičice („ringlice“) (/ fried onion rings). 
This includes the adjectives derived, for example, from the names of foreign cheeses: gorgonzolski, groerski, ementalski, parmezanski... (of or related to Gorgonzola, Gruy`ere, Emmental, Parmesan...) that are not present in RSANU (1959–2019). 
Ethnographic sources (Троjановић 1983) and (Ердељановић 1908) were carefully excerpted for the purpose of compiling the dictionary RSANU (1959–2019), but although around a hundred excerpted examples refer to the language of food, just a small number of them features the qualifier kuv.  
In the dictionary of foreign words (Клаjн and Шипка 2006) more attention has been dedicated to marking the lexemes belonging to the domain of cooking by using the qualifier kulin(arstvo) (culinary). 
The following definitions apply to some of the examples found in this dictionary: 
– musaka kulin. jelo od tanko izrezanog krompira, patlidžana ili tikvice sa iseckanim ili mlevenim mesom koje se peče u pećnici. tur. musaka od ar. musaqqa natopljen. (mousaka culin. A dish made of thinly sliced potatoes, eggplants or courgettes with chopped or minced meat fried in the oven. tur. musaka from ar. musaqqa soaked); 
– blanširati (...) kulin. kratko (pro)kuvati (povrće ili voće), (po)pariti. fr. blanchir izbeljivati. (to blanch (...) culin. to cook (vegetables or fruit), steam for a short period of time fr. blanchir.). 
But, there too, certain food ingredients or cookware do not have the marker kulin as in the examples: 
– paradajz (...) bot. jestiva povrtarska biljka, Solanum lycopersicum, crveni patlidžan, rajˇcica. nem. Paradiesapfel rajska jabuka (tomato (...) bot. edible vegetable plant, Solanum lycopersicum, crveni patlidžan /red tomato, rajčica / tomato. ger. Paradiesapfel apple from paradise); 
– šerpa (...) niska okrugla posuda u kojoj se čuvaju namirnice. Nem. Scherbe (pot (...) shallow, round dish for storing food. Ger. Scherbe). 
The missing qualifier kuv/kulin, and the way of defining the elements of the language of food do not prevent an intelligent dictionary reader from understanding the meaning of a certain word, based on its definition. 
However, such instances of the lack of precision are, most likely a serious obstacle, both in terms of dictionary formalization and automatic extraction of a sub-dictionary describing the language of food. 
A query that does not refer to the presence of a certain entry or its forms in the dictionary, but to other markers, such as qualifiers, for example cannot yield a reliable result. 
A comprehensive catalogue of the vocabulary of cooking could also be found in multilingual dictionaries aimed at the hospitality industry, e.g. (Banićević and Popović 2010; Jovanović 1970). 
These dictionaries are primarily intended for compiling multilingual menus and they, as a rule, consist of lists featuring names of dishes in two or more languages. 
These lists, obtained by excerpting different sources provide an insight into the equivalent dish names in different languages, especially dishes belonging to the international cuisine. 
But, such dictionaries offer no information about the sources of the entries selected to be featured in them, which makes their actual use in (Serbian) language questionable in certain cases. 
For instance, in (Banićević and Popović 2010), we come across exotic entries like čorba od kengurovog repa (kangaroo tail broth) (67) or konsome od peraja ajkule (shark fin consomm´e) (72). 
In special cases, the entry is accompanied by main ingredients, as in the example corbaˇ sa dronjcima (masa od brašna i jaja zakuvana u vrući bujon) dough (flake broth – a mixture of flour and eggs poured into a hot bouillon) (69). Thus, in the case of numerous entries, it turns out to be hard or impossible to get a clue as to their meaning based on dictionary information. 
Still, this lexicographic effort offers an insight into the issue of transfer of culinary content from one language to another. 
For instance, the French dish cailles sous le cendre is translated into Serbian as prepelica pod sačom (175), and into English as quails baked in ash; therefore, the Serbian translation introduces a regional concept of preparing food under a large metal lid (sač) that is not present either in French or in English. 
Conversely, even when the Serbian equivalent is linguistically precise, it cannot convey the meaning of the original entry. 
So, the corresponding Serbian translation equivalent of the French term potage cultivateur (Banićević and Popović 2010, 79) is zemljoradnička čorba (farmer’s broth), but this French traditional cuisine recipe is not known in our region. 
The way in which the material for the above-mentioned Serbian language dictionaries is collected via partial excerption of selected sources, followed by a choice of a certain number (not all) hits obtained in this manner, raises the question of entry relevance. 
Namely, since the information regarding entry frequency is missing, all entries are of equal importance, at the same level. 
Moreover, as shown by the above examples, partial excerption results in a number of (potential) entries becoming excluded from the dictionary material. 
Finally, the very structure of the language of food evades such a description of its lexis. 
The name of a dish makes sense only if it is accompanied by a recipe detailing how it is to be prepared. 
The entries belonging to the language of food would have to include, in addition to a description of meaning, a wide array of information establishing their social and cultural status, temporal and territorial distribution, dietary comments, etc. 
These remarks make it clear that conceiving and compiling a cooking dictionary describing different aspects of the language of food, requires set-ting up a wider corpus that incorporates diverse sources, as well as defining a more complex microstructure of a dictionary article. 
Although the lexis of the language of food has been largely described in Serbian dictionaries, this description is indirect, as seen in the previous section and the information giving indications as to its evolution is missing. 
Particularly, the complex processes determining the nature of the collective taste, as part of the national identity cannot be described on the basis of the existing lexicographic descriptions. 
As an illustration of the complexity of this process, it is possible to make a comparison between contemporary notions about the content of the Serbian cuisine, for instance, from a corresponding Wikipedia article, with the range of dishes from older cookbooks and similar sources. 
Thus, the salad section in Wikipedia features eight “typical” salads that are part of the contemporary view of “traditional salad”, which are, however, still missing from Pata’s Cookbook in that shape or form. 
Besides, dictionaries do not provide data on the frequency of the individual concepts belonging to the language of food, which would be considered to be a reliable criterion of their being widely known (accepted). 
So, the above article on salads mentions trljanica, but the web or Serbian language corpora show that the widespread knowledge about this salad is insignificant compared to others, such as srpska (Serbian), šopska (Shopska – Bulgarian) or ruska salata (Russian Salad – Salad Olivier). 
Trljanica is, therefore, a concept i.e. local cuisine dish that the wider public is not familiar with.  
The things that define the concepts related to dish names are recipes for their preparation (as precise as possible). 
The definitions beginning with dish made with flour. . . or with milk. . . that are a regular occurrence in lexicographic descriptions give only the most generalized impression of the dish type. 
On the other hand, a recipe depends, among other things, on the time when it was devised, as well as on the wider social context in which it was applied. 
The gradual integration of the concept of German soup into the contemporary Serbian menu lasting almost a whole century can serve as an illustration of that. 
Starting from the first cookbook in Serbian by hieromonk Jerotej Draganović Драгановић (1855), where the clear beef soup recipe (Продановић-Младенов 1984, 359) is named O govedini (On Beef )  and Midžina’s cookbooks, where it is named goveda čorba, buljon (beef  broth, bouillon) and later bela goveda čorba (white beef soup) to Pata’s Cookbook, where it appears as goveda supa (beef soup), separate from govedi bujon (beef bouillon), one and the same concept, accompanied by similar recipes, changes its name, but not its significant position, as compared to other “liquid dishes” (as described in dictionaries). 
The advance of this dish from the Serbian cuisine in Austria-Hungary to other parts of Serbia was gradual and the term supa itself took long to become accepted in Serbia at the time.  
The question now is whether it would be possible to set up a corpus of the language of food which would allow monitoring different parameters determining the evolution of eating habits over the last 150 years. 
For each query, such a corpus would have to provide information about the time when a certain food-related concept appears or disappears, the manner in which it is propagated in different social strata, the changes in diet under the influence of technologies and external cultural influences, etc. 
Searching the corpus would provide a more precise insight into the changes of the language of food and the changes of the collective taste alike. 
When did a certain word with a certain meaning appear, when did it change its meaning, when did it disappear? 
What are the changes in its frequency in the period covered by the corpus? 
The sources for compiling this kind of corpus collected up to the present moment are heterogenous. 
They include different written sources available, among which the primary ones are as follows: 
– Cooking handbooks, published or in manuscript; 
– Ethnographic studies on food and drink; 
– Historical studies, especially private life histories; 
– Cooking and gastronomy textbooks; 
– Dictionaries of various kinds; 
– Cooking monographs; 
– Excerpts from literary works (including translations into Serbian) de-scribing eating habits; 
– Statistical representations; 
– Other information-rich contributions like menus, etc. 
It is clear that this kind of material must be properly prepared in advance, so as to make it possible to search the corpus using complex queries. 
This primarily applies to detailed metadata about the sources themselves and labelling relevant classes of named entities in them. 
For example, Троjановић (1983) describes customs present in a wide area, defined either by the common name of a territory (Serbia, Montenegro, Herzegovina, Macedonia, Slavonia, the Balkan peninsula, etc.) or a region indicator (Srem, Pirot District, Vasojević District, etc.) or very precisely, by stating the place where a certain custom was encountered (Pirot, Tetovo, Kladovo, Aleksinac, Rud-nik, Mostar, Zlatibor, etc.). 
It is clear that, in that case, a concept belonging to the language of food gets specified in two ways: by the source itself (Троjановић 1983) and the location where it was recorded (e.g. Vasojević District). 
The entities referring to measurements in recipes pose a particular problem themselves as a result of different unit systems, as well as the lack of precision in stating measures in cooking handbooks (Krstev et al. 2017). 
The roles of individual sources listed above are not one and the same. 
Recipe books list names of dishes and indirectly the ingredients that the featured dishes consist of. 
As a rule, they reflect eating habits of the middle class at a certain point in history where the recipes are often strongly influenced by foreign culinary skills. 
Still, these handbooks too include certain recipes typical of the traditional cuisine (for instance, in connection with the customs related to religious holidays). 
The numerous recipes found in such books provide no details as to their acceptance (in the population). 
An important piece of information given by the cooking handbooks are instructions on the way in which a dish is prepared at a certain point in time and that is, as pointed out above, the true definition of the dish name. 
Moreover, the language of these handbooks reveals how certain cooking concepts got their names in Serbian. 
For example, the term šnicla (steak) instead of komadić (small piece) (Поповић-Миџина 1891) or kriška (slice) (HrL 1878) appears as late as in the cookbook by Мирковић (1922), while the cooking method of pohovanje (frying in breadcrumbs or batter), which used to be subsumed under the more general term przenjeˇ (frying), is named as late as in (Рустановић 1938). 
Cookbooks also provide information about the range of foodstuffs available at a certain time (Витас 2017), as well as an insight into the problem of naming new ingredients. 
The organization of the structure of cooking handbooks in order to make performing complex searches of their contents possible is another question to consider. 
Namely, a cooking handbook can be viewed as a separate item illustrating the cuisine of a period, and also as a collection of separate, independent recipes. 
In the latter case, linking the same or similar recipes from different handbooks enables us to follow their evolution. 
The presence of a dish in different handbooks separated by long stretches of time is an indication of the level of its integration into the national cuisine. 
On the other hand, among the wide variety of recipes offered by such handbooks, some recipes appear only once: these dishes that may have resulted from the inspiration of their authors are singular occurrences and do not form part of the usual range of widely accepted meals. 
For example, the recipe pečeno prase u aspiku (roasted young pig in aspic) found only in (Мирковић 1922, 59)	is more of an epitome of a display of gargantuan excess than part of the standard cuisine. 
The recipes themselves feature a description of the necessary ingredients, preparation procedures, utensil requirements, etc., which paints a picture of the state of cooking abilities and skills at a certain period. 
On the other hand, these elements of the language of food too undergo changes from one handbook to another under the influence of new kinds of foodstuffs and food preparation techniques. 
For example, Jerotej Draganović’s cookbook still does not mention štednjak (cooker), while as early as the first edition of Midžina, some twenty years later, cooking with the help of a cooker goes without saying. 
The missing parameter in these sources is the confirmation of a concept independently from culinary or ethnographic descriptions. 
For example, among the myriad of broths, mentioned in the different editions of Midžina, some are amply confirmed in other sources, while in the case of others, there are no examples, except those included in these handbooks, showing that they were prepared. 
In (Драгановић 1855), there is a recipe for čorba od čokolade (chocolate broth) that is not present in later cookbooks. 
In Midžina (second edition) ostrige (oysters) are mentioned as many times as dunavski som (Danube catfish), therefore, we have reason to wonder whether, in the mid-19th century Novi Sad, oysters were as commonly eaten as catfish. 
The information about the frequency of a concept cannot be found in cookbooks either. 
This kind of data can be obtained indirectly in the extracts from the corpus of literary works or newspaper articles describing eating customs, as shown, for example, in (Vitas 2022) or in the rare surviving restaurant or individual feast menus, e.g. (Чолак-Антић 2004; Костић 2019). 
Monographic publications such as (Mijo 2012; Onfre 2002; Пелапрат 1971; Kostjukovič 2007), describing culinary habits in other places are of special importance to the extent to which they show how the content from other culinary cultures, whose numerous elements will be subsequently integrated into the Serbian eating repertory can be expressed in Serbian. 
Such foodstuffs or dishes get described in the translation of the foreign source first, and only later included in the Serbian menu. 
The gradual integration of new ingredients and dishes into what the local cuisine has to offer can be illustrated via the example of inclusion of sea fish and sea food into the Serbian menu repertory. 
In the third edition of her cookbook, Midžina mentioned morska riba (sea fish) remarking that she had never cooked them (Поповић-Миџина 1891, 149), while there was no mention of lignje (squids). 
The fourth significantly altered edition mentions see fish that come salted or dried “therefore half-prepared” (Поповић-Миџина 1911, 409), but without stating the individual fish varieties. 
In Rustanović’s cookbook (Рустановић 1938, 349) there are instructions for preparing sea fish that differentiate between blue, white and wild sea fish. 
The latter group also includes ligni (squid) or kalamari (calamari) in addition to lignica, hobotnica (octopus) and sipa (cuttlefish). Moreover, the cookbook features recipes for their preparation. 
This cookbook features orada (sea beam) too, “one of the tastiest kinds of sea fish”, which is also referred to as zlatna (gold – Gilt-head bream) in the picture on the page 350. 
In Pata’s cookbook the selection of sea fish is limited tosardele (sardines) and skuša (mackerel), but the author says that “sea fish is regularly available in our open-air markets” (Марковић 1959, 93). 
Sea bream is missing from her book, but the list of sea fish includesbrancin (sea bass) (47). 
It mentions kalamari (calamari) (or olignji and ulinje) as well. 
A vast selection of sea fish would appear in (Ненадовић 1978) (zubatac (common dentex), brancin (sea bass), orada (sea bream), cipol (flathead grey mullet), murina (moray...), together with kalamari (calamari). 
Morski rakovi (crayfish) that used to be present in earlier cookbooks, no longer appear in newer books on cooking, while the concept of morski plodovi (sea food) is not recorded in any of the analyzed cooking handbooks. 
This example shows how the food that does not originate from the local sources gets gradually integrated into the national menu through the development of the appropriate lexis. 
The described corpus of texts on cooking provides a foundation for constructing a lexical base depicting the lexicon of the sublanguage of food in Serbian. 
In view of the content of the corpus, different types of dictionaries describing certain aspects of this sublanguage can be exported from such a base (dictionary of foodstuffs, historical dictionary describing the evolution of culinary terms, multilingual dictionary, no-fat cookbook, etc. (Stanković, Krstev, et al. 2018; Stanković, Stijović, et al. 2018). 
In an earlier stage of processing of the texts belonging to the corpus, it is necessary to solve, above all, the issue of a high degree of variability of the names used, especially in older texts, describing the same concept, referring either to the ingredients, pre-made products or preparation procedures and the necessary utensils. 
This problem can be solved either by establishing an appropriate semantic network (Vujičić-Stanković, Krstev, and Vitas 2014), or via a collection of graphs (in a manner in which they are used in the Unitex system, for example) whose content depicts the content of individual nodes of the network assigning to it a canonical node representative (as a transducer output). 
Thus, plant-based ingredients have different names, depending on author and time of appearance. 
Sargarepa (carrot), whose name was taken from Hungarian is mentioned in cooking handbooks under the following names as well: 
– žuta repa, šargarepa (Драгановић 1855; RK 1915; Мирковић 1922); 
– žuta  repa  –  šargarepa,  mrkva  (žuta  repa),  žuta  mrkva  (šargarepa) (Поповић-Миџина 1891, 1911); 
– žuta repa (HrL 1878); 
–	mrkva and šargarepa (Марковић 1959; AFZ 1952; Ненадовић 1978; 	Продановић-Младенов 1984); 
– mrkva (Jokić et al. 1983); 
–	mrkva, šargarepa, karota (Рустановић 1938); 
–	šargarepa (Стоjанић 2004; Рашић 1999). 
In addition to these, in (Лазић 2000) there are also the names žuta zelen, šangarepa, crvena mrkva. 
The beginnings of use of this plant in the Serbian cuisine in the early 20th century are discussed in (Костић 2006): 
“Naši seljaci [...] nisu gajili tada spanać, grašak, paradajz, salatu, šargarepu i ostalu zelen. Sve su to zemunske piljarice donosile za kuće u kojima je već vladala zapadnjaˇcka kuhinja.” (“Our peasants [...] weren’t growing spinach, peas, tomatoes, lettuce, carrots or other vegetables at that time. All that was brought by the green grocers in Zemun for the households that had already been dominated by the western cuisine.“) (397) 
In the Eltec-corpus there is only one mention of žuta repa (yellow turnip), while šargarepa and mrkva are absent. 
According to the corpus of the contemporary Serbian language, šargarepa (with the frequency 667), is the dominant term nowadays, mrkva (56) appears more rarely, and karota (3) and žuta repa (1) are quite rare. 
By bringing together these different names for šargarepa (carrot) and other regional ones recorded in RSANU (1959–2019) (besides Latin Daucus carota hortensis, there are also mrkvela, mrkuca, merlen, etc.) to form a node in the semantic network (or a node in a Unitex graph) it is possible to neutralize the variability of names in the course of indexing or performing searches of the corpus. 
The described semantic network construction would make neutralization of lexical variability possible, thus simplifying solving the next complex task – identification of identical (or similar) recipes. 
The content of morphological electronic dictionaries allows recognizing a certain number of culinary concepts, as described in (Крстев and Лазић 2015; Vujičić-Stanković, Krstev, and Vitas 2014), via semantic markers associated with the dictionary entries. 
For the purposes of indexing the corpus it would be necessary to broaden and refine the set of markers in order to specify particular features of entities. 
For example, the marker Zool describes zoological terminology, while the marker Anim describes raw foodstuffs, but these markers cannot be used to automatically distinguish foodstuffs of animal origin the way it is done in cooking handbooks (kinds of meat, fish, poultry, etc.) 
The recipes featured in cooking handbooks are either variations of an already well accepted dish or occasional dishes that are not repeated in other places and whose existence has not been confirmed elsewhere. 
Thus, for instance, pileći paprikaš (chicken stew) or govedi gulaš (beef goulash) are present in almost all above-mentioned books on cooking. Therefore, they can be said to belong to the range of national dishes, despite their Hungarian origins, while the recipe for patka u malaginom umokcu (duck in Malaga dip) is featured only in early editions of Midžina, disappearing as early as in the fourth edition. 
Some dishes disappear in time, or undergo changes and others get integrated into the recognizable range of dishes that make up the national cuisine. 
Mrka or crna čorba (dark or black broth), for example, present in early books on cooking, used to be an important dish at feasts that was gradually reduced to a modest bone soup, eventually disappearing from the repertory of the described dishes altogether. 
A recognizable element of the national cuisine today, ražnjići (shish-kebab) appears for the first time, as late as in (Поповић-Миџина 1911). 
The decision regarding the concepts forming part of the traditional repertory of dishes can be made by using the frequencies found in the different corpora of the contemporary Serbian language SrpKor13 and the ELTeC-corpus as a starting point. 
Thus, pasulj (beans), pečenje (roast), paprikaš (stew), gulaš (goulash), ćevapčić or ćevap (grilled minced meat fingers ), đuvec (saut´eed rice with meat and vegetables), sarma (stuffed cabbage leaf rolls), sarmica (small stuffed leaf rolls), podvarak (stewed sauerkraut), kapama (lamb stew), ajvar (red pepper chutney), etc. are among the high frequency dishes in the first of the two corpora that are confirmed in the other. 
Some of these terms have undergone a change of meaning over time. 
Certain old dishes have disappeared, e.g. janija (lamb stew) and papazjanija (beef and pork stew) while new ones have appeared: pljeskavica (grilled meat patties), šnicla (steak). The same applies to the ingredients used, e.g. susam (sesame). 
The structure of the dictionary articles in the database should be conceived in such a way so as to make accumulating different kinds of information related to the entries possible. 
An example to be emulated in that respect is the lexicon (Zirojević 2019) where the featured foodstuffs are illustrated not only by lexicographic, but also various other information, including etymological, historical and nutritional data. 
The selection of entries for a dictionary of this kind would have to incorporate, in addition to the foodstuffs and dishes present in the Serbian cuisine, culinary techniques and equipment. 
A description of dishes in such a database can be accompanied not only by linguistically relevant data, but also by outstanding examples of recipes (including their development over time), as well as chosen, primarily, literary examples that often define the social, cultural and geographic context in which a certain dish appeared. 
The neglected sublanguage of food in Serbian, as described in this paper, presents a multifaceted research challenge. 
Possibly the most comprehensive rationale for such an undertaking was given by Escoffier in his introduction to the famous Larousse gastronomique (Montagn´e 1949; Courtine 1986): 
“The history of the table of a nation is a reflection of the civilization of that nation. To show the changes in the order and serving of meals from century to century, to describe and comment on the progress of the French cuisine, is to paint a picture of the many stages through which a nation has evolved since the distant times.“  
The available dictionaries, private life histories, ethnographic writings, books on cooking and other available sources provide only well-known, accessible parts of a complex mosaic in which the Serbian gastronomy developed. 
On the other hand, the dishes that are nowadays considered to be an integral part of the Serbian traditional table conceal the continuum of events determining the current dominant local taste. 
The road traveled by today’s “traditional” dish, from a regional one and even more often under pressure of cultural influences from elsewhere has largely remained hidden. 
The basis that would finally ensure the provision of the material for a “good and comprehensive dictionary of cooking” is the described concept of a corpus of the language of food. 
In view of the contemporary dictionary form, it is clear that it will emerge and get constructed as a particular, possibly multimedia, database before any paper version. 
Its excerpts can yield specific or general dictionaries of cooking, but also, besides that, such information technology structures make possible continuous updating and development of the inner structure by adding new content (Stanković, Krstev, et al. 2018). 
Apart from the dictionary, as an occasional overview of the state of the database, the opportunity for long-term monitoring of the eating habits and customs would allow to get a clearer insight into what is indeed part of the usual – traditional? - eating habits of the local population. 
The applications of this kind of database are wider than lexicography. They can be perceived in other research areas, especially in culture studies, as well as historiographical and ethnographic studies, along with the everyday practice of the hospitality industry and tourism. 
