Introduction

The success of modern intelligent software systems designed for processing texts in natural language requires a large experimental linguistic base [4; 6; 7; 9; 10; 11; 16], which is what this study focuses on. The demand for corpus data corresponds to the emergence of appropriate technical capabilities [1; 3; 6; 14; 15; 17; 21; 22; 25; 31].

Corpora are usually designed for multiple use by many users, so their markup and linguistic support must be unified in some way [20; 23; 24; 28]. Corpora standards usually address the compatibility of markup types.

Difficulties in developing natural language are also predetermined by certain external factors, including the presence of hundreds of natural languages, with typical set of syntactic rules. Depending on the context, words possess multiple meanings within a single language [18; 19; 29]. Complications become obvious even at the level of symbols, digits and graphemes. As machine-readable language, Text Coding Initiative rationalizes texts processing in terms of stressing valid information as found within textness and textuality features [2; 13; 26]. To add, since 1980 Text Coding Initiative [12] has been a global group of written language researchers working in the field of digital humanities.

Resulting from the joint efforts of researchers, the TEI documents such a standard in the TEI Guiding Principles. These recommendations are flexible and fully adaptable to diverse projects; using TEI would advocate the compatibility of the chosen project within other non-digitalized humanitarian projects.

The reason is that for more than two decades TEI has been widely used in projects in a variety of disciplines, fields and subjects internationally. The constant interaction of the wide user community through the TEI organization confirms the importance of the text coding standard and signals about its continuous development and validity.

Within the Translation Studies research, saving and preserving texts in electronic format provides undoubted vistas: high speed information retrieval, easily-conducted editing procedures, multimedia adaptability and localization, hyperlinks options. Scanned texts can be stored in various graphic forms. However, for more efficient use, scanned texts as converted to digital text formats are of paramount advantage. Digital text format reduces the file size and allows other applications to reformat, search, or process text within the dimension of modern digitalized society and academia.

TEI parameterization of Erich Fromm's novel

Justification

The relevance of this study is proven by its theoretical and practical significance for research both in linguistics and in related disciplines: textual linguistics, translation studies, and literary studies. Since the main focus is on fiction text coding with the consequent practical approbation of the results obtained, this study proposes a method of automated search for suggested tags based on sentence selection for both original English material and its Ukrainian translation. Prioritized in the current study is the method of TEI application in relation to a bilingual fiction text.

We opt for (a) describing and characterizing modern TEI methods, (b) applying them to the practical markup of English and Ukrainian texts and (c)preparing materials for concordance on the basis of the results obtained within the scope of tasks conducted.

With this in mind, we concentrate on: 1) cursory analyzes of the main theoretical provisions of corpus linguistics and linguistic markup;

2) preliminary study of the original and translation of the analysed novel;

3) markup process of the original and translation of the novel in accordance with the recommendations of the TEI; 4) creation a sample of sentences, both in the original and in the translation, containing at least one tag from the set created by us; 5) scrutiny of the features, liminations and the definition of a set of tags; 6) choice of information technology tools for creating our software; 7) organizing tags for English and Ukrainian data; 8) analyzes of the results using the created software.

The study was based on the original and the translation of Erich Fromm's novel -The Art of Loving‖ [27; 32].

The scientific novelty of this work is that the study of the proposed set of tags and sentences in the original and translation has never been automated. The software created helps to distinguish convergent and divergent features between marking up textness peculiarities of the original and the translation, correspondingly.

Mandatory elements of TEI

The Text Encoding Initiative is the standard for presenting textual material in digital form using text encoding tools. As we have already mentioned, in this paper work we rely on the instructions of TEI, as well as the XML markup language [8]. Due to its advantages and wide popularity, XML has become the metalanguage of choice for expressing the rules of descriptive text encoding in TEI. The advantage of TEI is obvious since TEI is expressed in terms of extensible markup language (XML) and it provides all the procedures and mechanisms to adapt to specific project needs.

The TEI guidelines define an open standard that can normally be applied to any text and may satisfy any purpose. So, as we have already mentioned, in our work we rely on TEI instructions, as well as XML markup language.

Any XML encoding language consists of five components:

• XML language declaration; • Elements;

• Attributes; • Links to objects;

• (P) CDATA.

XML language declaration

XML documents have to enclose with an XML language declaration that indicates the version of XML used. In the markup of both the original novel and the translation, we also cited XML language declarations. For example, <?xml version="1.0" encoding="UTF-8"?> The question mark "?" in the declared language of XML signals that this is the so-called instruction for processing this document. The following entry shows that there is an XML that complies with the recommendations of version 1.0; UTF-8 or Unicode character encoding is being used. An XML declaration is optional; however, in case of usage it can only appear at the beginning of an XML file [26].

Another important counterpart of the TEI documents is the title of the TEI. The TEI guidelines state that such a virtual "identity card" is an important part of every TEI document. It must be encoded in the <teiHeader> element before the actual text content in the <text> part. The minimum TEI header for our research looks as follows [26].

Elements and attributes

The two-paragraph document below is an example of a part of an XML document that represents both information and meta-information. Information (plain text) is contained in XML elements separated by opening tags (such as <div>) and closing tags (such as </div>). Additional information about these XML elements can be provided in attributes that consist of a name (for example, @ n) and a value meaning (for example, "I"). Attributes can only be found in the element's initial (opening) tag. XML comments are separated by opening tags (<! -) and closing tags (->). Everything in the comments is ignored by the XML processing software: it is said to be "commented on".

Figure 4: Information and metainformation tagging

The initial and most commonly used element of structural markup is a paragraph. Paragraphs may contain any other elements that we will consider below, as well as those elements that are specific to certain types of texts. It is customary to designate a paragraph by the element p. Examples of the markup of the original and translation of Erich Fromm's novel The Art of Loving: This excerpt from the original and the translation of the novel demonstrates how other elements of the phrase level (in this case s element representing the sentence) can be nested in paragraphs, but not vice versa. Momentous aspect of coding of primary data in the corpus is also the interpretation of editorial markup, which provides additional information about the text, the specifics of lexical items, possibly about its author, etc. Editorial markup usually equals to highlighting certain elements of the text. Visually, the selection (or highlighting) can be conveyed in various ways: say it, by means of bold type and / or italics in lowercase, bold type and / or italics in uppercase, petit, letter spacing, etc. Applying the TEI principle for an uninterpreted typographic selection we use the <hi> element, which marks a word or phrase that is graphically different from the main text and the reason for this selection is unknown or unidentified. The <hi> element uses the rend attribute to specify the type of text fragment selection.

The following text with the highlighted words ‗object,‗ faculty, ‗love should be coded as follows: Another example illustrates the situation when tags can be nested; also when not only one word can be highlighted, but an entire expression or sentence (the needed textness feature prioritized). However, most paragraphs do not have a clear internal structure, but contain prose being coded as a combination of symbols, references to the essence, phrases, marking. They may also contain built-in elements such as lists, figures or tables. Though, since our markup is made within the framework of a prosaic creative work, there are only a few inserted poetic passages. For example,

Figure 13: Poetic lines formatting

In the Ukrainian translation this passage is presented as follows:

Figure 14: Poetry in translated-product

There are also foreign words and phrases in the analyzed novel that we encode with the <foreign> element. We use the @xml: lang attribute to define foreign languages. Consider the following examples.

Link (reference) to objects

Another issue with corpus markup is character (symbol) encoding. In previous versions of SGML TEI, special English characters (such as the £ pound), European-accented characters (such as é), and other non-ASCII (US Standard Code for Information Sharing) characters were replaced with links to objects which are separated by "&" and ";". The object name is a line between these two characters.

For example, an entity reference for a "less than" sign < is "&lt;" and an entity reference for an ampersand is "&amp;", "&pound;" which means £, "& eacute;" means é. Object references can be useful in the following cases: symbolic data that cannot be easily typed on the keyboard or not used in XML, shielding reserved characters in XML, providing "template text", i.e. text that is used or can be reused in new contexts or programs. However, since our paper is based on the marking up of the fiction novel, neither in the original nor in the translation are there any special symbols that we have marked in this way.

Although declaring object references can solve the problem of representing special characters (symbols) for languages that have only a small number of such characters, this is not a rational solution for languages that contain many special characters. Therefore, Unicode, which has adopted XML as a mandatory character set for all XML documents, is an attempt to solve this problem.

Unicode is truly multilingual in terms of displaying characters from a very large number of writing systems, and therefore promises to provide a standard for multilingual encoding of corpus characters.

XML Editor allows us to directly enter characters in most writing systems around the world and saves those characters in a way that can be transferred directly between different computer systems. XML support has made it possible to spread these markup schemes more widely in a multilingual context [5].

(P)CDATA

All text in an XML document is usually parsed. When an XML element is parsed, the text between the XML tags is also parsed. The text data that the XML parser will parse is called parsed character data or PCDATA.

An XML document often contains data that does not need to be parsed by an XML parser. For example, characters such as "<" and ">" are not used in XML elements because the parser interprets them as the beginning or end of new elements, resulting in an error message. We can avoid this by looking at them through links to "& lt;" and "& amp;" or CDATA in the document.

The CDATA section starts with <! [CDATA [and ends with]]>. Everything inside the CDATA section is considered by the parser as plain text.

Marking of 'love' textness by means of TEI

Having analyzed the original and translation of the novel, we also identified a number of special concepts that dominate Fromm's work, namely -motherly love, fatherly love, brotherly love, erotic love, love for God and other types of love. Based on these concepts, a number of tags have been created that correspond to them -<mlove>, <flove>, <blove>, <elove>, <slove>, <Glove>, <olove>.

Let us consider examples of sentences with the concept of maternal love.

Figure 15: Sample of taglining

Therefore, the main function of markup is to present textual material in digital form using text encoding. In order for research to be easily shared, analyzed, reused and stored, texts on humanities field must be coded according to a standard that defines common vocabulary, grammar and syntax. As we can see, these recommendations are fully adapted and adjustable to a specific project, while improving the compatibility of this project with other projects using TEI.

Creation and practical use of "Concordance 'Art of Loving'"

In recent years, Python has become increasingly popular due to its efficiency in development areas such as Machine Learning and Data Science. It is so due to its scalability and flexibility, which is so necessary in this segment of programming. The popularity of Python accredited thanks to the fact that this language is able to solve a wide range of problems and its application is possible on all popular platforms today, except for encoding mobile applications. So you can highlight the following benefits of Python:  Possibility of expansion -there are libraries and frameworks for any type of tasks and needs;  Simplicity of syntax -everything superfluous has been removed from syntax, the code is pure and clear without superfluous brackets and expressions;  Interpretability -the Python interpreter exists for all popular platforms;  PEP -the only standard for writing code that makes code supported and readable even when switching from one programmer to another;  Open Source -Python interpreter code is open, which allows anyone interested in language development to participate in its development and improvement. Creating a program for a concordance based on the original and translated novel "The Art of Love" by Erich Fromm is useful, because it will greatly simplify our further translation studies research. Also, taking into account all the advantages of the Python language, the dictionary was created in this programming language. Therefore, an integrated development environment for the PyCharm version 3. In the window on the left we see a field with a list of sentences that contain <mlove>, <flove>, <blove>, <elove>, <slove>, <Glove>, <olove> tags. When you first start the program, the window on the right is empty. However, if you enter a word or tag in the search field (Fig. 17) and click the ‗Search' button, the search results will be displayed in the right window (Fig. 18). When you click on a sentence in the right window, the translation of this sentence will be displayed, as in Fig. 17 In addition to tags, you can search by ordinary words or phrases; then the search result will be those sentences that contain these words (Fig. 20). In case of the absence of such words or phrases, the window on the right remains empty (Fig. 21). Apart from this, our program has a second window called "Statistic". This window displays certain statistics. For example, if you enter any of the tags in the search field, the "Statistic" window will display statistics for this tag, namely the number of sentences in which this tag is used (Fig. 22). The graph also shows a separate column "others", which contains sentences that contain other tags in addition to the <mlove> tag. For a better understanding, the following tips have also been developed, given in the column (Fig. 23). They show the number of sentences shown on the graph. In turn, the "Help" button opens a document with system help of using Python (Fig. 25).