Problem areas

Manuscripts being what they are, it sometimes happens that parts of them can no longer be read at all or with the usual degree of confidence. The markup of such sections – whether they be long or short – is discussed here.

The relevant tags are:

• <unclear>
• <damage>
• <gap>
• <space>
• <supplied>

Unclear text

Sometimes a letter, a word, a phrase or even a whole passage cannot be transcribed with certainty because we can’t work out what the element or elements in question are. The <unclear> tag marks up such text.

"Unclear" text in the corpus may be the product of one of three processes:

• the ink may have faded
• someone may have retraced text
• someone may have tried to amend the original

These three processes are encoded in the reason attribute.

The following examples demonstrate the <unclear> tag in use:

<w><unclear reason="faint">a</unclear>ual</w>
<w><unclear reason="retraced">a</unclear>ual</w>
<w><unclear reason="correction">a</unclear>ual</w>

Damage

Probably the most frustrating experience of a transcriber is when he has to admit defeat: so great is the degree of damage to his text that nothing whatsoever can be read. One of the reasons for such damage is that a hole has developed in the manuscript. But not all holes are the product of wear, tear and rodents. Some were present when the scribe began his task.

To our mind, the difference between a naturally occurring hole and one which developed after the text was written is important enough to merit separate tags: there may be loss of information if damage occurred after the text was written but there is no information loss if the damage was pre-existing. We have, however, complied with the TEI guidelines and encoded both possibilities with the catch-all <damage /> tag.

The types of damage which we have encountered fall into three groups. These are encoded with the type attribute:

• natural hole: typically found in a piece of skin through which passed one of the legs of its original owner; the text continues the other side of the hole
• unnatural hole: due to a variety of reasons, e.g. water, mice or other pests; some text may have been lost
• cropped: an unfortunate consequence of subjecting the manuscript to a binder’s knife; some text will have been lost

In the transcriptions, therefore, complete <damage /> tags have one of following three formats:

<damage type="natural hole" />
<damage type="unnatural hole" />
<damage type="cropped" />

The TEI guidelines allow for describing various other attributes of a damaged area, such as its size and who or what was responsible for it. Of particular relevance to our project was the size of a damaged area, which could be defined with an extent attribute. However, to have accurately measured these spaces would not only have been costly and time-consuming, but would also have detracted us from our primary task – to transcribe the texts. We therefore leave such matters to others.

Gaps

<gap /> indicates a point where we’ve been obliged to omit material in a transcription because the text is illegible.

As with <damage />, the TEI guidelines list several attributes which may be used in conjunction with <gap />. Again we decided that the most practicable course of action for the present project was to forgo the details so that we could devote our main energies to transcribing.

Spaces

Sometimes a scribe has left an atypical amount of space in a text. This happens most often when we have a gap which was clearly intended for a word or an ornate initial capital which was never added. The presence of such ‘significant’ space is encapsulated by <space />.

Example:

<w><space />ri</w><w>gwr</w><w>doeth</w>

Editorially supplied text

<damage> typically and <gap> always encode parts of a manuscript which for some reason are no longer there or have become illegible. But the existence of damage does not mean that all is lost. Comparing a text with other related texts may enable us to be reasonably confident that we know what has been lost. Text between <supplied> signifies such text.

The reason attribute indicates why the text has had to be supplied:

• binding: the text has been swallowed up by the binding
• cropped: the text has been cut, probably by a binders knife
• damage: the text has been damaged
• illegible: the text cannot be read
• sic: the text is missing because of scribal error

Two typical examples would be:

<w><supplied reason="illegible">y</supplied></w>
<w><supplied reason="binding">y</supplied></w>

Supplied text is typically the product of team consenus. We have not ‘claimed’ such additions by putting our names to them. Sometimes, however, supplied text is the product of a previous researcher. The provenance of such additions is indicated by the resp attribute. In the Peniarth 46 manuscript, for example, the resp attribute records that the text was supplied by J. Gwenogvryn Evans:

<w><supplied reason="illegible" resp="Gwenogvryn">y</supplied></w>