The following element is used to identify textual subdivisions in the un-numbered style: As a member of the class att.typed, this element has the following additional attributes:

Using this style, the body of a text containing two parts, each composed of two chapters, might be represented as follows:

The following elements are used to identify textual subdivisions in the numbered style: As members of the class att.typed these elements all bear the following additional attributes:

The largest possible subdivision of the body is div1 element and the smallest possible div7. If numbered divisions are in use, a division at any one level (say, div3), may contain only numbered divisions at the next lowest level (in this case, div4).

Using this style, the body of a text containing two parts, each composed of two chapters, might be represented as follows:

Within the same front, body, or back element, all hierarchic subdivisions must be marked using either nested div elements, or div1, div2 etc. elements nested as appropriate; the two styles must not be mixed.

The choice between numbered and un-numbered divisions will depend to some extent on the complexity of the material: un-numbered divisions allow for an arbitrary depth of nesting, while numbered divisions limit the depth of the tree which can be constructed. Where divisions at different levels should be processed differently (for example to ensure that chapters, but not sections, begin on a new page), numbered divisions slightly simplify the task of defining the desired processing for each level, though this distinction could also be made by supplying this information on the type attribute of an un-numbered div. Some software may find numbered divisions easier to process, as there is no need to maintain knowledge of the whole document structure in order to know the level at which a division occurs; such software may, however, find it difficult to cope with some other aspects of the TEI scheme. On the other hand, in a collection of many works it may prove difficult or impossible to ensure that the same numbered division always corresponds with the same type of textual feature: a chapter may be at level 1 in one work and level 3 in another.

Whichever style is used, the global n and xml:id attributes (section ) may be used to provide reference strings or labels for each division of a text, where appropriate. Such labels should be provided for each section which is regarded as significant for referencing purposes (on reference systems, see further section ).

As indicated above, the type and subtype attributes provided by the att.typed class may be used to provide a name or description for the division. Typical values might be book, chapter, section, part, or (for verse texts) book, canto, stanza, or (for dramatic texts) act, scene. The following extended example uses numbered divisions to indicate the structure of a novel, and illustrates the use of the attributes discussed above. It also uses some elements discussed in section and the p element discussed in section . Book I. Of writing lives in general, and particularly of Pamela, with a word by the bye of Colley Cibber and others.

It is a trite but true observation, that examples work more forcibly on the mind than precepts: ...

As an alternative (or complement) to this use of the type attribute to characterize neutrally named division elements, the modification mechanisms discussed in section may be used to define new elements such as chapter, part, etc. To make this simpler, a single member model class is defined for each of the neutrally named division elements: model.divLike (containing div), model.div1Like (containing div1), model.div2Like (containing div2), etc. For example, suppose that the body of a text consists of a series of diary entries, each of which is potentially divided into entries for the morning and the afternoon. This might be represented in any of the following ways. First, using the un-numbered style:

More information about the customization techniques exemplified here is provided in .

In most situations, the textual subdivisions marked by div or div1 (etc.) elements will be both complete and identically organized with reference to the original source. For some purposes however, in particular where dealing with unusually large or unusually small texts, encoders may find it convenient to present as textual divisions sequences of text which are incomplete with reference to the original text, or which are in fact an ad hoc agglomeration of tiny texts. Moreover, in some kinds of texts it is difficult or impossible to determine the order in which individual subdivisions should be combined to form the next higher level of subdivision, as noted below.

To overcome these problems, the following additional attributes are defined for all elements in the att.divLike class:

For example, an encoder might choose to transcribe only the first two thousand words of each chapter from a novel. In such a case, each chapter might conveniently be regarded as a partial division, and tagged with a div element in the following form:

The head element is used to identify a heading prefixed to the start of any textual division, at any level. A given division may contain more than one such element, as in the following example: Etymology (Supplied by a late consumptive usher to a grammar school)

The pale Usher — threadbare in coat, heart, body and brain; I see him now. He was ever dusting his old lexicons and grammars, ...

Unlike some other markup schemes, the TEI scheme does not require that headings attached to textual subdivisions at different hierarchic levels have different identifiers. All kinds of heading are marked identically using the head tag; the type or level of heading intended is implied by the immediate parent of the head element, which may for example be a div1, div2, etc., an un-numbered div, or any member of the model.listLike class. However, as with div elements, the encoder may choose to extend the model.headLike class of which head is the sole member to include other such elements if required.

In certain kinds of text (notably newspapers), there may be a need to categorize individual headings within the sequence at the start of a division, for example as main headings, or detail headings: this may readily be done using the type or subtype attribute. Specific elements are provided for certain kinds of heading-like features, (notably byline, dateline, and salute; see further section ), but the type or subtype attributes must be used to discriminate among other forms of heading. These attributes are provided, as elsewhere, by the att.typed attribute class of which the head element is a member.

In the following example, taken from a British newspaper, the lead story and its associated headlines have been encoded as a div element, with appropriate model.divTop elements attached:

In older writings, the headings or incipits may be longer than in modern works. When heading-like material appears in the middle of a text, the encoder must decide whether or not to treat it as the start of a new division. If the phrase in question appears to be more closely connected with what follows than with what precedes it, then it may be regarded as a heading and tagged as the head of a new div element. If it appears to be simply inserted or superimposed—as for example the kind of pull quotes often found in newspapers or magazines, then the quote, q, or cit element may be more appropriate.

The trailer element, which can appear at the end of a division only, is used to mark any heading-like feature appearing in this position, as in this example:

In addition to headings of various kinds, divisions sometimes include more or less formulaic opening or closing passages, typically conveying such information as the name and address of the person to whom the division is addressed, the place or time of its production, a salutation or exhortation to the reader, and so on. Divisions in epistolary form are particularly liable to include such features. Additional elements for the detailed encoding of personal names, dates, and places are provided in chapter . For simple cases, the following elements should be adequate:

The byline and dateline elements are used to encode headings which identify the authorship and provenance of a division. Although the terminology derives from newspaper usage, there is no implication that dateline or byline elements apply only to newspaper texts. The following example illustrates use of the dateline and signed elements at the end of the preface to a novel:

Where a sequence of such elements appear together, either at the beginning or end of an element, it may be convenient to group them together using one of the following elements: The following examples demonstrate the use of the opener and closer grouping elements:

For further discussion of the encoding of dates and of names of persons and places, see section and chapter .

The argument element may be used to encode the prefatory list of topics sometimes found at the start of a chapter or other division. It is most conveniently encoded as a list, since this allows each item to be distinguished, but may also simply be presented as a paragraph. The following are thus both equally valid ways of encoding the same argument:

An epigraph is a quotation from some other work, a saying, or a motto, appearing on a title page, or at the start of a division. It may be encoded using the special-purpose epigraph element, as in the following example: E. M. Forster Howards End Only connect... When an epigraph contains a quotation, this may often be associated with a bibliographic reference. In such cases, it is recommended additionally to group the quotation and its source together using the cit element, as in the following example:

For discussion of quotations appearing other than as epigraphs refer to section .

A postscript is a passage added after the signature of a letter or, less frequently, the main portion of the body of a book, article, or essay. In English a postscript is often abbreviated as P.S. or PS, and postscripts are often introduced by labels with one of these abbreviations, as in the following example.

Other than elements from the model.divWrapper, model.divTop, or model.divBottom classes, every textual division (numbered or un-numbered) consists of a sequence of ungrouped macro.component elements (see ). The actual elements available will depend on the modules in use; in all cases, at least the component-level structural elements defined in the core will be available (paragraphs, lists, dramatic speeches, verse lines and line groups etc.). If the drama module has been selected, then other component- or phrase- level items specialized for performance texts (for example, cast lists or camera angles) will be available, as defined in chapter ) will be available. If the dictionary module is in use, then dictionary entries, related entries, etc. (as defined in chapter ) will also be available; if the module for transcribed speech is in use, then utterances, pauses, vocals, kinesics, etc., as defined in chapter will be available; and so on.

Where a text contains low-level elements from more than one module these may appear at any point; there is no requirement that elements from the same module be kept together.

Examples of composite texts which should be represented using the group element include anthologies and other collections. The presence of common front matter referring to the whole collection, possibly in addition to front matter relating to each individual text, is a good indication that a given text might usefully be encoded in this way; this structure may be found useful in other circumstances too.

For example, the overall structure of a collection of short stories might be encoded as follows: The Adventures of Sherlock Holmes First published in between July 1891 and December 1892 Adventures of Sherlock Holmes Adventure I. — A Scandal in Bohemia By A. Conan Doyle.

To Sherlock Holmes she is always the woman. ...

A text which is a member of a group may itself contain groups. This is quite common in collections of verse, but may happen in any kind of text. As an example, consider the overall structure of a typical collection, such as the edition of Crashaw's poetry. Following a critical introduction and table of contents, this work contains the following major sections: (a collection of verse first published in 1648) (a second collection, published in 1652) (a third collection, published in 1648) I (a collection of fragments all taken from a single manuscript) II (a further collection of fragments, taken from a different manuscript)

Each of the three collections published in Crashaw's lifetime has a reasonable claim to be considered as a text in its own right, and may therefore be encoded as such. It is rather more arbitrary as to whether the two posthumous collections should be treated as two groups, following the practice of the edition. An encoder might elect to combine the two into a single group or simply to treat each fragment as an ungrouped unitary text.

The edition reprints the whole of each of the three original collections, including their original front matter (title pages, dedications etc.). These should be encoded using the front element and its constituents (on which see further section ), while the body of each collection should be encoded as a single group element. Each individual poem within the collections should be encoded as a distinct text element. The beginning of the whole collection would thus appear as follows (for further discussion of the use of the elements div and lg for textual subdivision of verse, see section and chapter ): The poems of Richard Crashaw Edited by J.R. Tutin

The group element may be used in this way to encode any kind of collection of which the constituents are regarded by the encoder as texts in their own right. Examples include anthologies or collections of verse or prose by multiple authors, florilegia, or commonplace books, journals, day books, etc. As a fairly typical example, we consider , an anthology edited by Paul Fussell and published in 1987 by W. W. Norton. This work comprises the following major sections: Front matter (title page, acknowledgments, introductory essay) The Beginnings The Eighteenth Century and the Grand Tour The Heyday Touristic Tendencies Post Tourism Back matter (permissions list, index) Each titled section listed above comprises a group of extracts or complete texts from writers of a given historical period, preceded by an introductory essay. For example, the second group listed above contains, inter alia, the following: Prefatory essay Five letters by Lady Mary Wortley Montagu An extract from Swift's Two poems by Alexander Pope Two extracts from Boswell's Journal A poem by William Blake Each group of writings by a single author is preceded by a brief biographical notice. Some of the extracts are quite lengthy, containing several chapters or other divisions; others are quite short. As the above list indicates, the texts included range across all kinds of material: verse, prose, journals and letters.

The easiest way of encoding such an anthology is to treat each individual extract as a text in its own right. A sequence of texts by a single author, together with the biographical note preceding it, can then be treated as a single group element within the larger group formed by the section. The sequence of single or composite texts making up a single section of the work is likewise treated, together with its prefatory essay, as a single group within the work. Schematically: The Beginnings The Heyday

Note that the editor's introductory essays on each author may be treated as texts in their own right (as the essays on Lady Mary Wortley Montagu and Alexander Pope have been treated above), or as front matter to the embedded text, as the essay on Swift has been. The treatment in the example is intentionally inconsistent, to allow comparison of the two approaches. Consistency can be imposed either by treating the Swift section as a group containing one text by Swift and one by the editor, or by treating the Montagu and Pope sections as text elements containing the editor's essays as front matter. Marked in the second way, the Pope section of the book would look like this:

The essays on The Eighteenth Century and the Grand Tour and other larger sections could also be tagged as front matter in the same way, by treating the larger sections as text elements rather than group elements.

Where, as in this case, an anthology contains different kinds of text (for example, mixtures of prose and drama, or transcribed speech and dictionary entries, or letters and verse), the elements to be encoded will of course be drawn from more than one module. The elements provided by the core module described in chapter should however prove adequate for most simple purposes, where prose, drama, and verse are combined in a single collection.

For anthologies of short extracts such as commonplace books, it may often be preferable to regard each extract not as a text in its own right but simply as a quotation or cit element. The following component-level elements may be used to encode quotations of this kind: For example, the chapter of extracts which appears in the front matter of Melville's might be encoded as follows:

An important characteristic of the unitary or composite text structures discussed so far is that they can be regarded as forming what is mathematically known as a tesselation covering the whole of the available text (or text division) at each hierarchic level. Just as an XML document has a single root element containing a single tree, each node of which forms a properly nested sub-tree, so it seems natural to think of the internal structure of a text as decomposable hierarchically into subparts, each of which is a properly nested subtree. While this is undoubtedly true of a large number of documents, it is not true of all. In particular, it is not true of texts which are only partly tesselated at a given level. For example, if a text A is contained by text B in such a way that part of B precedes A and part follows it, we cannot tesselate the whole of B. In such a case, we say that text A is a floating text.

The floatingText element is a member of the model.divPart class, and can thus appear within any division level element in the same way as a paragraph. For example, texts such as the or the might be regarded as containing many floating texts embedded within another single text, the framing narrative, rather than as groups of discrete texts in which the fragments of framing narrative are regarded as front or back matter.

As an example, we consider an 18th century text , by Jane Barker (1726). This lengthy narrative contains nearly a hundred distinct tales embedded (as the title suggests) in a single patchwork. The work begins by introducing the central character, Galecia, but within a few pages launches into a distinct narrative, the story of Captain Manly:

Galecia one Evening setting alone in her Chamber by a clear Fire, and a clean Hearth ... reflected on the Providence of our All-wise and Gracious Creator....

She was thus ruminating, when a Gentleman enter'd the Room, the Door being a jar... calling for a Candle, she beg'd a thousand Pardons, engaged him to sit down, and let her know, what had so long conceal'd him from her Correspondence.

To my parents, Ida and Max Fish

The answer this book gives to its title question is there is and there isn't.

Chapters 1–12 have been previously published in the following journals and collections: chapters 1 and 3 in chapter 10 in (1980) . I am grateful for permission to reprint.

The remembrance of the many wrongs offred thee, and thy unreproued vertues, adde greater sorrow to my miserable state, than I can utter or thou conceiue. ... ... yet trust I in the world to come to find mercie, by the merites of my Saiuour to whom I commend thee, and commit my soule.

Once more, gentle reader I appeal unto thee, from the shameful ignorance of the Editor, by whom Our own Specimen of Virgil hath been mangled in such miserable manner, that scarce without tears can we behold it. At the very entrance, Instead of προλεγομενα, lo! προλεγωμενα with an Omega! and in the same line consulâs with a circumflex! In the next page thou findest leviter perlabere, which his ignorance took to be the infinitive mood of perlabor but ought to be perlabi ... Wipe away all these monsters, Reader, with thy quill.