This chapter introduces markup targeting at digital scholarly editions of music.
Critical Apparatus, the alignment of multiple sources / witnesses of the same musical text is discussed.
Editorial Markup covers editorial observations in and interventions to the text.
Genetic Markup finally deals with the special requirements and needs of genetic editions in music.
11.1. Critical Apparatus
This chapter describes how to encode differences between multiple exemplars of the same musical work (often referred to in MEI as ‘sources’). The mechanisms and elements described in this chapter are closely related to their counterparts in the TEI guidelines. It is also important to refer to chapter
Editorial Markup of these guidelines, especially concerning the
choice element described therein.
11.1.1. General Usage
The following elements are defined in the critApp Module:
(reading) – Contains a single reading within a textual variation.
app element always encapsulates the differences between varying sources. Therefore, it must contain at least two child elements. Possible child elements are
rdg, which use the same model, but have a different meaning: Whereas
lem is used for prioritizing one alternative, a
rdg has no such additional meaning and simply indicates a reading as found in one or more sources. Accordingly,
lem is allowed only once in
rdg may appear as often as necessary.
lem) elements use the @source attribute to point to one or more descriptions of the bibliographic sources containing the material they mark:
The @seq attribute may be used on
rdg to record the sequence of a series of readings. In the following example, the material in source B is marked as sequential to (and perhaps derived from) the reading in source A:
If a source has additional content that is not found in other sources, an empty
rdg element may be used to indicate the lack of material in the other sources. In the following example, source 1 includes material that is not found in sources 2 and 3:
When working with a large number of sources, it might seem tedious to provide references for all sources. However, use of the
rdg element without @source is not recommended because such an encoding is not explicit and is therefore difficult to process.
11.1.2. Variants in Musical Content
app element may be used to accommodate textual variation at nearly any point in a musical text. For example, it may be used to indicate minor differences such as stem directions:
or to indicate more significant differences, such as the insertion of extra measures:
However, the flexibility in the location of
app places a burden on the encoder to ensure that the
lem elements are used correctly; that is, the content of every
lem has to be a valid replacement for its parent
app, even though this cannot be controlled effectively by the MEI schema.
11.1.3. Variants in Score Definitions
In addition to its use for differentiation of the musical content of multiple sources,
app may also be utilized to describe the layout of different scores, even when the musical content itself remains the same. An example of this is two sources that have the same content, but a different ordering of staves on which the content is written. By definition, the order of staves in MEI is described in and derived from the order of
staffDef elements in
scoreDef, not from the order of
staff elements within a
staff element in a measure points to its corresponding
staffDef using the same value for @n on both elements.
This rather loose mechanism makes it possible to point dynamically to the correct staff definition for a given source. The following example demonstrates how this can be accomplished for two sources, both presenting a two-staff score, but with differing staff order. No further
app element is necessary within the
measure to describe the alternative score order of the sources.
When unique values for @n on
layer are provided, it is possible to reallocate layers in the same fashion as staves.
This mechanism may also be used to describe not only differing page orientations, formats and margins, but also clefs and keys.
The use of
app in conjunction with
staffDef illustrates the greater flexibility of connecting
staffDef by a shared @n value. A technically more robust alternative to @n would be to use the @def attribute on
staff, which points to the @xml:id of a
staffDef. However, this strong connection would be tied to one specific
staffDef, and would not allow to pick one alternative out of an
11.1.4. Nesting Apparati
In some situations, musical sources will agree at one level while differing at a lower level. For these cases,
app elements may be nested to any level necessary. In the following example, there are three sources, two of which agree on the addition of a measure, but differ in the content of the added measure:
app elements, it is important that the value(s) in the child
rdg element’s @source attribute must be a strict subset of the ancestor
rdg element’s @source value.
11.2. Editorial Markup
It is often necessary to render an account of any changes made to a musical text during its creation (and any subsequent editing) and to accommodate editorial comment necessitated by an editorial process. The elements and attributes described in this chapter may be used to record such editorial interventions, whether made by the composer, the copyists of the manuscript, the editor of an earlier edition used as a copy text, or the current encoder/editor.
The scope of the elements described herein is therefore the description of features relating to the genesis, later revision and editorial interpretation of a text. Mechanisms for describing multiple sources are described in chapter
Critical Apparatus of these Guidelines, while the full setup for genetic editions is described in chapter
The elements described in this chapter may be contained by a wide range of other MEI elements and, in turn, may contain a variety of elements. The encoder must assume responsibility for the appropriateness of the markup; that is, a great many combinations of editorial and transcriptional markup are technically possible, but care must be taken to see that the encoding does not contravene the rationale of these Guidelines. In general, it should be ensured that a file would be valid if the editorial markup would be omitted, as such a validation cannot be ensured in an efficient way by the MEI schema.
For most of the elements discussed here, some encoders may wish to indicate both a responsibility; that is, a coded value indicating the person or agency responsible for making the editorial intervention in question, and an indication of the degree of certainty which the encoder wishes to associate with the intervention. The elements discussed here thus may potentially carry the following optional attributes:
Many of the elements discussed here can be used in two ways. Their primary purpose is to indicate that their content represents an editorial intervention (or, in some cases, the lack of intervention) of a specific kind. Sometimes, pairs or other meaningful groupings of such elements can be recorded, then wrapped within the special purpose
Groups a number of alternative encodings for the same point in a text.
Wrapping elements this way enables the encoder to represent, for example, a text in its ‘original’, uncorrected form alongside the same text in one or more ‘edited’ forms. Making use of this style of representation, software may dynamically switch between the ‘Urtext view’ of the text and one or more ‘views’ of the text after the application of the encoded editorial interventions.
Three categories of editorial intervention are discussed by the remainder of this chapter:
indication or correction of apparent errors;
indication of regularization of variant, irregular, non-standard, or eccentric forms; and
editorial additions, suppressions, and omissions.
MEI offers methods for marking abbreviations in prose, as in the following example:
or abbreviations in the music itself, as in the following example:
The generic @type attribute may be used to classify the abbreviation according to a convenient typology. Sample values include:
suspension: the abbreviation provides the first letter(s) of the word or phrase, omitting the remainder;
contraction: the abbreviation omits some letter(s) in the middle;
brevigraph: the abbreviation comprises a special symbol or mark;
superscription: the abbreviation includes writing above the line;
acronym: the abbreviation comprises the initial letters of the words of a phrase;
title: the abbreviation is for a title of address (Dr, Ms, Mr, …);
organization: the abbreviation is for the name of an organization;
geographic: the abbreviation is for a geographic name.
This tag is the mirror image of the
expan tag (not to be confused with the
expansion element described in
Content of Musical Divisions). Both
expan allow the encoder to transcribe an abbreviation and its expansion. In the case of
abbr, the original is transcribed as the content of the element and the expansion as an attribute value, while
expan reverses this. The choice between the two is up to the user. For example:
abbr tag is not required; if appropriate, the encoder may transcribe abbreviations in the source text silently, without tagging them. If abbreviations are not transcribed directly but expanded silently, then the MEI header should indicate this is the case. The @cert attribute signifies the degree of certainty ascribed to the expansion of the abbreviation. The @expan attribute gives an expansion of the abbreviation. The @resp attribute contains an ID reference to an element containing the name of the editor or transcriber responsible for supplying the expansion of the abbreviation.
When the content of the @abbr or @expan attributes requires additional markup, an attribute cannot be used. In this case, the abbreviated and expanded forms must be presented within elements. Furthermore, as alternatives to each other, the
expan elements must be wrapped by the
choice element, as described above. The previous example, where the ‘o:’ in ‘pno:’ is written as superscript, would be encoded as:
Many musical scores make use of various kinds of shorthand notation which omit some parts of the score that have already been written elsewhere. Typical examples for this are symbols that indicate repetition of the preceding measure or beat. In MEI, these symbols can be encoded using the
beatRpt elements respectively. Often, similar graphical symbols (often one or two slashes, “//”) are used to mean that the current staff should have the same or similar content as another staff.
colla parte directives have a less strictly-defined scope than the ‘Rpt elements’ (
multiRpt). That is, rather than specifying the repetition of content of a particular duration, like a measure or beat, colla parte instructions can refer to material of any length. In order to encode such scribal shorthand, MEI offers the
cpMark element, which allows filling of blank spaces in the score with horizontally and/or vertically distant material.
(copy/colla parte mark) – A verbal or graphical indication to copy musical material
Like any other ‘controlEvent’ (see
Events and Controlevents),
cpMark is placed in the score using the @staff and @tstamp attributes. The end point of the mark itself, when necessary, may be indicated using the @tstamp2 attribute. The source material, which is intended to be inserted in the space indicated by the copy mark, can be identified by the attributes @origin.tstamp, @origin.tstamp2, @origin.staff and @origin.layer. While @origin.tstamp provides the relative distance from the beginning of the “gap”, @origin.tstamp is relative to the position identified by @origin.tstamp. However, @origin.tstamp defaults to the same value as @tstamp2 and should only be provided when necessary. When neither @origin.staff nor @origin.tstamp are provided, they take the same values as the cpMark’s @staff and @tstamp attributes; that is, they indicate a strict ‘vertical’ or ‘horizontal’ copy.
In the example above, there are no less than three different copy instructions, which need to be encoded with four
cpMark elements. First, Weber inserts characters from “a” to “f” in red ink to identify filled measures. Then, he repeats the same characters in empty measures, which indicates that the content from the filled measures should be copied here. While one could try to encode this with just one
cpMark element, it is both clearer and easier to process when using two elements.
The second and third shorthand indications are written in the second violin (lower staff). Here, Weber writes “unis.[ono]”, silently omitting the reference to the first violin. His next shorthand (“in 8va”) additionally instructs the copyist to double the written material in another octave. This information can be captured using the @dis and @dis.place attributes on
Text used as a copy mark, like the letters in the Weber example, may be encoded as content of the
cpMark element. In the case of non-text marks, the @altsym and @facs attributes may be used to refer to a graphical surrogate.
Depending on the purpose of the encoding, the omitted parts in the score may be filled with
mSpace elements of appropriate duration or silently overwritten with the content that the
cpMark identifies. Also, these two options may be combined through the use a
choice element whose
expan children explicitly encode a transcription of the original ‘gap’ (in abbr) and the result of the insertion of the indicated material (in expan, see
11.2.2. Apparent Errors
When the source material to be encoded is manifestly faulty, an encoder or transcriber may elect simply to correct it without comment, although for scholarly purposes it will often be more generally useful to record both the correction and the original state of the text. The elements described here enable all three approaches, and allows the last to be done in a way that makes it easy for software to present either the original or the correction.
(correction) – Contains the correct form of an apparent erroneous passage.
The following examples show alternative treatment of the same material. The text to be encoded contains a chord (c4, e4, g4, a4), where c4, e4, and a4 are quarter notes, but g4 is incorrectly written as a half note.
An encoder may choose to silently correct the engraver’s error:
or the correction may be made explicit:
Alternatively, the encoder may simply record the typographic error without correcting it, either without comment or with a
sic element to indicate the error is not a transcription error in the encoding:
If the encoder elects to record the original source text and provide a correction for the sake of transparency, both
corr may be used, wrapped in a
choice element. The order of the
corr elements is not significant:
An indication of the person or agency responsible for the emendation can be provided as follows:
Here the @resp attribute has been used to indicate responsibility for the correction. Its value (#editTrans.JK) is an example of the pointer values discussed in section
Pointers and References. In this case, the @resp attribute points to a
name element within the metadata header, but any element might be indicated in this way, if the correction has been taken from some other source. The @resp attribute is available for all elements which are members of the
att.common class (through
att.edit class makes available a @cert attribute (through
att.evidence), which may be used to indicate the degree of editorial confidence in a particular correction, as in the following example:
Where, as here, the correction takes the form of amending information present in the text being encoded, the encoder should use the
corr element. Where the correction is present in the text being encoded, and consists of some combination of visible additions and deletions, the elements
add and / or
del should be used. For additional information on the use of
del, see section
Additions and Deletions. Where the correction takes the form of an addition of material not present in the original because of physical damage or illegibility, the
supplied element may be used. Where the ‘correction’ is simply a matter of expanding abbreviated notation, the
expan element may be used.
11.2.3. Regularization and Normalization
When the musical source makes extensive use of unusual symbol shapes or non-standard notation features, it may be desirable for a number of reasons to regularize it; that is, provide ‘standard’ or ‘regularized’ forms that are equivalent to the non-standard forms.
As with other such changes to the source text, the changes may be made silently (in which case the MEI header should still specify the types of silent changes made) or may be explicitly marked using the following elements:
Groups a number of alternative encodings for the same point in a text.
Consider this traditional soprano clef appearing somewhere in the course of a musical piece:
An encoder may choose to preserve this original clef, but flag it as nonstandard from the perspective of current practice by using the
orig element with no attributes specified, as follows:
Alternatively, the encoder may indicate that the clef has been modernized into a G-clef by using the
reg element with no attributes specified, as follows:
As another alternative, the encoder may encode both the old and modernized shapes, so that applications may render both at the reader’s will:
As described above, the @resp attribute may be used to specify the agent responsible for the regularization.
11.2.4. Additions, Deletions, and Omissions
The following elements are used to indicate when single notational symbols have been omitted from, added to, or marked for deletion from, a musical text. Like the other editorial elements described in this chapter, they allow for a wide range of editorial practices:
(deletion) – Contains information deleted, marked as deleted, or otherwise indicated as
superfluous or spurious in the copy text by an author, scribe, annotator, or corrector.
22.214.171.124. Omissions, Unclear Readings, Damage, and Supplied Readings
Encoders may choose to omit parts of the source for reasons ranging from illegibility, (making transcription difficult or impossible), to editorial policy, e.g., systematic exclusion of poetry or prose from an encoding. The full details of the policy decisions concerned should be documented in the MEI header (see section
Encoding Description). Each place in the text at which omission has taken place should be marked with a
gap element, optionally with further information about the reason for the omission, its extent, and the person or agency responsible for it, as in the following examples:
Note that the extent of the gap may be marked precisely using attributes @unit and @extent.
Unlike TEI, MEI does not offer a desc element for further description of the reason for a gap. Instead, an
annot may refer to the gap via its @startid, @endid, or @plist attributes and provide additional information.
unclear element is used to mark passages in the original which cannot be read with confidence, or about which the transcriber is uncertain for other reasons, as for example when transcribing an illegible source. Its @reason and @resp attributes are used, as with the
gap element, to indicate the cause of uncertainty and the person responsible for the conjectured reading.
Where the difficulty in transcription arises from an identifiable cause, the @agent attribute signifies the causative agent. The @cert attribute signifies the degree of certainty ascribed to the transcription of the text contained within the
unclear element. Where the difficulty in transcription arises from action (partial deletion, etc.) assignable to an identifiable hand, the @hand attribute may record the hand responsible for the action.
When the reason for a gap in the encoding is damage of the document carrier (the paper on which the document is written, for example), the
damage element should be used instead of the
gap element. In the case of damage resulting from an identifiable cause, the @agent attribute signifies the causative agent. The @degree attribute signifies the degree of damage according to a convenient scale. A
damage tag with this attribute should only be used where the text may be read with some confidence; data supplied from other sources should be tagged as
supplied. The @extent attribute indicates approximately how much text is in the damaged area, in notes, measures, inches, or any appropriate unit, where this cannot be deduced from the contents of the tag. For example, the damage may span structural divisions in the text so that the tag must then be empty of content. In the case of damage (deliberate defacement, etc.) assignable to an identifiable hand, the @hand attribute signifies the hand responsible for the damage.
Sometimes the editor provides information not present in the source material. These conjectures or emendations are marked up in MEI using the
The following example demonstrates the use of the
supplied element in combination with
When the presumed loss of text arises from an identifiable cause, @agent signifies the causative agent. When the presumed loss of text arises from action (partial deletion, etc.) assignable to an identifiable hand, the @hand attribute signifies the hand responsible for the action. The @reason attribute indicates why the text has to be supplied, e.g. ‘overbinding’, ‘faded ink’, ‘lost folio’, ‘omitted in original’, etc. The @source attribute contains the source of the supplied text. The editor(s) responsible for supplied material may be recorded in the @resp attribute. The value of @resp must point to one or more identifiers declared in the document header. The @cert attribute signifies the degree of certainty ascribed to the supplied material.
126.96.36.199. Additions and Deletions
del elements may be used to record where material has been added or deleted in the source material.
The following example demonstrates the usage of
add to mark up a note being added to an existing chord:
The next example shows how
del may be used to capture the information that two measures have been cancelled. As seen in this example, the @rend attribute is used to specify the method of deletion.
Additional information for both elements may be specified using attributes. Whereas the @hand attribute marks responsibility for the textual change, the @resp attribute is used to refer to the editor who identified this textual change as such. The @cert attribute signifies the degree of certainty ascribed to the identification of the hand of the deletion or addition.
add element should not be used to mark editorial changes, such as supplying a note omitted by mistake from the source text or a passage present in another source. In these cases, either the
supplied tags should be used instead.
188.8.131.52. Substitutions, Restorations, and Handshifts
When several interventions to the musical text are to be regarded as a single action, they may be grouped using the
subst element. The most common combination is a replacement of portions of the musical text using both the
del element, as seen in the following example:
An intervention closely related to substitution is the restoration of a previously deleted section. For this purpose MEI offers the
restore element, which may contain a
del or other content directly.
The following example illustrates an instance where a lyric was cancelled and later restored by overwriting it:
The @desc attribute gives a prose description of the means of restoration. The @cert attribute signifies the degree of certainty ascribed to the identification of the hand of the restoration. The @type attribute may be used to indicate the action cancelled by the restoration. The @resp attribute contains an ID reference to an element containing the name of the editor or transcriber responsible for identifying the hand of the restoration. The @hand attribute signifies the hand of the agent which made the restoration.
MEI offers a
handShift milestone element that can be used to mark a change of scribe or scribal style.
The @character attribute describes characteristics of the hand, particularly those related to the quality of the writing, e.g., ‘shaky’, ‘thick’, regular’. A description of the tint or type of ink, e.g., ‘brown’ or the writing medium, e.g., ‘pencil’, may be placed in the @medium attribute.
The new hand may be identified using the @new attribute, while the previous hand may be recorded in the @old attribute. The @resp attribute contains an ID reference to an element containing the name of the editor or transcriber responsible for identifying the change of hand. The @cert attribute signifies the degree of certainty ascribed to the identification of the new hand.
When using this element within a layer, it is important to ensure that all layers and staves are considered. Every
handShift affects only the content of its own layer and staff, even in the following measures. Therefore, there must be a separate
handShift for every
layer. This mechanism allows the description of shifts at timestamps that differ between each staff.
11.3. Genetic Markup
Genetic editions try to trace the creation of a (musical) work in all its recorded details, from the first sketches to the ‘final’ complete text. The aim of genetic textual criticism is to investigate compositional working and thinking processes - the genesis of compositions. In contrast to traditional scholarly editions, which focus on the constitution of a performable text of a work, Genetic Textual Criticism focuses on the process of production, the gradual elaboration of musical thoughts while writing. It is dependent on the availability of comprehensible traces of these writing processes. Genetic editions often have to deal with significant uncertainties, and they require a considerable amount of markup, as detailed below.
11.3.1. Encoding Genetic States
Leaving aside temporary breaks, a compositional process is continuous: the composer’s writing operations have happened in a strict order, which could be specified if his working would have been filmed. However, this exact order is rarely ever recoverable from the surviving mansucripts, prints or other materials. Instead, relative statements can be made – the red pencil must have been written after the brown ink etc. Instead of a continuous movie, scholars are often only able to reconstruct a slide show, reflecting certain recoverable states of the composition. Very often, those states have a local range only – the order of two states on one page may be known, as is the order of two other states on a second page. This doesn’t necessarily allow to identify the succession of all four states.
MEI utilizes the
genDesc element to describe the recoverable genetic states of a work. It is nested inside
music and may contain a number of
Describes a distinctive state in the textual development of a work.
A genetic description is used to bundle all known states of a work. The @ordered attribute may be used to specify whether the order of child elements of the
genDesc matches their temporal order, or if their temporal order is unknown. As
genDesc may be self-nested, it is possible to specify the order of some states within a larger unordered set of genetic states, or to position a set of states with unknown temporal order in a larger ordered set.
In the above example, the temporal relation of states A, B and all of C is not known, but it is known that C1 precedes C2 and C3.
Even when the temporal order of a set of states is not fully recoverable, some arguments about relative chronology may be available. It is possible to encode these statements with the precision the editor feels comfortable with, utilizing the attributes @next and @prev (for immediate successors / predecessors), and @precedes and @follows (for relative successors / predecessors).
In the example above, state X of the encoded work is established by the change from a C clef to a G clef. Other states preceding state X will read a C clef, while state X and succeeding states read a G clef. A genetic state of the work is constituted by performing all textual operations referencing that state, i.e. by carrying out all additions, deletions and restorations.
The @instant attribute on
del etc. allows to specify that corresponding modification has been carried out immediately after writing the original text, so that no separate genetic state has been established.
It is up to encoder to identify the appropriate level of granularity: In an ideal world, the writing or cancellation of a single note would constitute a new state. In practice, this level of detail isn’t feasible, and the combination of multiple writing operations into larger logical operations seems inevitable. However, this may range from very large tasks (‘replacing the second movement of a symphony’) to very small ones (‘adding the slurs for the viola in measures 22 and 23’), depending on the intentions and scope of the encoding.
11.3.3. Encoding Metatexts
The arguments used to establish a chronological order of genetic states are sometimes found in external sources like letters, but very often they are to be found in the witnesses holding the musical text, even though they are typically not part of the text itself. Examples for such arguments are the writing medium, spacing, marginal notes, among others.
Some of these so-called ‘metatexts’ can be encoded using MEI, namely those that are written into the relevant sources. For this purpose, MEI offers the
metaMark element, as known from the TEI.
A graphical or textual statement with additional / explanatory information about the
musical text. The textual consequences of this intervention are encoded independently via
other means; that is, with elements such as <add>, <del>, etc.
A metaMark is provided as a ‘controlEvent’ (see
Events and Controlevents); as such, by convention, it should be encoded at the end of the
measure where it first occurs. It is highly recommended to specify the function of the metaMark using its @function attribute, which may take the following values:
confirmation: confirmation of a previous textual decision; i.e., cancellation of a deleted passage in a different writing medium.
addition: denoted material is to be inserted in the musical text.
deletion: denoted material is no longer part of the musical text.
substitution: denoted material is replaced, either by the musical text pointed at with the @target attribute or the musical content of the metaMark element itself.
clarification: attempt to clarify a potentially illegible or otherwise unclear part of the musical text.
question: marks a section of the musical text which is to be considered further.
investigation: marks a section of the musical text as an investigation of the consequences of certain compositional decisions or potential alternatives.
restoration: declares a formerly cancelled part of the musical text as valid again.
navigation: clarification of the reading order of the musical text.
Some metaMarks may have actual content, like marginal notes. This content may be transcribed inside the
metaMark element. It also has a @facs attribute to refer back to the corresponding sections of a facsimile.
It is important to keep in mind that
metaMark elements do not encode the textual consequences they transport – this is an encoding of the sign, not of its meaning, which can be encoded in other elements like
The above excerpt from a Beethoven manuscript holds a number of different metaMarks, some of which are encoded in the following examples:
The metaMark above captures the word ‘gut’ (good) Beethoven wrote below the red pencil, which indicates that the formerly deleted text of the two measures shown shall be kept.
metaMark transcribes Beethoven’s marginal note explaining the same situation as above.
This third metaMark covers one of the letters Beethoven inserted to clarify the pitch of that given note.
11.3.4. Genetic Changes at the Page Level
In genetic editions, it may also be of interest to trace when pages are added and / or removed from manuscripts. The general information about pages can be encoded using the
foliaDesc element, as described in
Description of folia. It is possible to wrap the elements used there, including
cutout with editorial markup like
del. These elements can then be used to reference
genState as described in
Referencing Genetic States.