Skip to content

Commit

Permalink
extended section on NEs by pseudonymisation example using gap and sup…
Browse files Browse the repository at this point in the history 
…plied
  • Loading branch information
@luengen
luengen committed Dec 11, 2023
1 parent 786d5de commit 5a4e803
Showing 1 changed file with 86 additions and 66 deletions.
152 changes: 86 additions & 66 deletions P5/Source/Guidelines/en/CMC-ComputerMediatedCommunication.xml
View file
Original file line number Diff line number Diff line change
Expand Up @@ -1161,78 +1161,98 @@ See the file COPYING.txt for details.
</p>
</div>

<div xml:id="CMCnames">
<head>Named entities</head>

<p>Named entites (NEs) may be marked up using <gi>name</gi> or the elements encoding
different subcategories of names as described in <ptr target="#ND"/>. In the following
chat example (adapted from <ptr target="#BIB_DCK"/>), nicknames are linked to a
<gi>person</gi> entry as shown in Section <ref target="#CMCParticipants"/> via the
<att>ref</att> attribute.
<egXML xmlns="http://www.tei-c.org/ns/Examples" xml:lang="de" source="#BIB_DCK">
<!-- Dortmunder Chatkorpus, chat 2213001 -->
<post modality="written" generatedBy="system" rend="color:black" synch="#f2213001.t007"
type="standard" who="#f2213001.A04" xml:id="f2213001.m27.eg35">
<name ref="#f2213001.A04" type="NICK">
<w lemma="Konstanze" pos="NE" xml:id="f2213001.m27.t1">Konstanze</w>
</name>
<w lemma="versuchen" pos="VVPP">versucht</w>
<name ref="#f2213001.A03" type="NICK">
<w lemma="Nasenloch" pos="NN">nasenloch</w>
</name>
<w lemma="die" pos="ART">den</w>
<w lemma="Wunsch" pos="NN">wunsch</w>
<w lemma="zu" pos="PTKZU">zu</w>
<w lemma="erfüllen" pos="VVINF">erfüllen</w>
<!-- ... -->
</post>
</egXML>
</p>
<div xml:id="CMCnames">
<head>Named entities</head>

<p>Named entites (NEs) may be marked up using <gi>name</gi> or the elements encoding different
subcategories of names as described in <ptr target="#ND"/> such as <gi>surname</gi> or
<gi>geoName</gi>, or <gi>rs</gi> for a general referencing string. In the following chat
example (adapted from <ptr target="#BIB_DCK"/>), nicknames are linked to a <gi>person</gi>
entry as shown in Section <ref target="#CMCParticipants"/> via the <att>ref</att> attribute.
<egXML xmlns="http://www.tei-c.org/ns/Examples" xml:lang="de" source="#BIB_DCK">
<post modality="written" generatedBy="system" rend="color:black" synch="#f2213001.t007"
type="standard" who="#f2213001.A04" xml:id="f2213001.m27.eg35">
<name ref="#f2213001.A04" type="NICK">
<w lemma="Konstanze" pos="NE" xml:id="f2213001.m27.t1">Konstanze</w>
</name>
<w lemma="versuchen" pos="VVPP">versucht</w>
<name ref="#f2213001.A03" type="NICK">
<w lemma="Nasenloch" pos="NN">nasenloch</w>
</name>
<w lemma="die" pos="ART">den</w>
<w lemma="Wunsch" pos="NN">wunsch</w>
<w lemma="zu" pos="PTKZU">zu</w>
<w lemma="erfüllen" pos="VVINF">erfüllen</w>
<!-- ... -->
</post>
</egXML>
</p>

<p>In the the following version of the same chat snippet, the text strings with the nicknames
have been replaced by category label strings for the purpose of anonymisation. <egXML
xmlns="http://www.tei-c.org/ns/Examples" xml:lang="de" source="#BIB_DCK">
<post modality="written" generatedBy="system" rend="color:black" synch="#f2213001.t007"
type="standard" who="#f2213001.A04" xml:id="f2213001.m27.eg35">
<name ref="#f2213001.A04" type="NICK">
<w pos="NE" xml:id="f2213001.m27.t1">
<gap reason="anonymisation" unit="token" quantity="1"/>
<supplied reason="anonymisation">[_FEMALE-PARTICIPANT-A04_]</supplied></w>
</name>
<w lemma="versuchen" pos="VVPP">versucht</w>
<name ref="#f2213001.A03" type="NICK">
<w pos="NN">
<gap reason="anonymisation" unit="token" quantity="1"/>
<supplied reason="anonymisation">[_PARTICIPANT-A04_]</supplied></w>
</name>
<w lemma="die" pos="ART">den</w>
<w lemma="Wunsch" pos="NN">wunsch</w>
<w lemma="zu" pos="PTKZU">zu</w>
<w lemma="erfüllen" pos="VVINF">erfüllen</w>
<!-- ... -->
</post>
</egXML>
</p>
<p>In the preceding example, pairs of a <gi>gap</gi> and a <gi>supplied</gi> element encode
the fact that some substring has been removed and replaced with another string for
anonymisation purposes. Note that in the example, the <gi>name</gi> and the <gi>w</gi>
elements and their attributes also provide some categorial information about what has been
removed. Using <gi>gap</gi> and <gi>supplied</gi> to record the anonymisation is especially
recommendable when the original name or referencing string has been
<soCalled>pseudonymised</soCalled> i.e. replaced by different referencing string of the
same ontological category (such as replacing the female name <hi rend="italic"
>Konstanze</hi> by the female name <hi rend="italic">Kornelia.</hi>) In that case, the
markup would be the only place where it can be seen that a pseudonymisation has been carried
out as in the following version of the example.</p>
<p>
<egXML xmlns="http://www.tei-c.org/ns/Examples" xml:lang="de" source="#BIB_DCK">
<post modality="written" generatedBy="system" rend="color:black" synch="#f2213001.t007"
type="standard" who="#f2213001.A04" xml:id="f2213001.m27.eg35">
<name ref="#f2213001.A04" type="NICK">
<w pos="NE" xml:id="f2213001.m27.t1">
<gap reason="pseudonymisation" unit="token" quantity="1"/>
<supplied reason="pseudonymisation">Kornelia</supplied>
</w>
</name>
<w lemma="versuchen" pos="VVPP">versucht</w>
<!-- remainder of the post -->
</post>
</egXML>
</p>
</div>


<p>In the the following version of the same chat snippet, the text strings with the
nicknames have been replaced by category label strings for the purpose of anonymization.
The category string and the <gi>name</gi> encoding offer some information about what has
been removed.
<!-- %%%%%%%%% still working on how to anonymize as of 2023-12-08 %%%%%%%%% -->
<egXML xmlns="http://www.tei-c.org/ns/Examples" source="#BIB_DCK">
<post modality="written" generatedBy="system" rend="color:black" synch="#f2213001.t007"
type="standard" who="#f2213001.A04" xml:id="f2213001.m27.eg36">
<name ref="#f2213001.A04" type="NICK">
<w pos="NE">
<gap reason="anonymised" unit="word" quantity="1"/>
</w>
<w pos="NE">
<gap reason="anonymised" unit="word" quantity="1"/>
</w>
</name>
<w lemma="versuchen" pos="VVPP">versucht</w>
<rs ref="#f2213001.A03" type="anonymised">
<w pos="NN">[_PARTICIPANT-A03_]</w>
</rs>
<w lemma="die" pos="ART">den</w>
<w lemma="Wunsch" pos="NN">wunsch</w>
<w lemma="zu" pos="PTKZU">zu</w>
<w lemma="erfüllen" pos="VVINF">erfüllen</w>
<!-- ... -->
</post>
</egXML>
</p>
</div>

<div xml:id="CMCmultimodal">

<head>Multimodal CMC</head>

<p>As explained in Section <ptr target="#CMCUnits"/> the
elements <gi>post</gi>, <gi>u</gi>, <gi>kinesic</gi>, and
<gi>incident</gi> are available to to encode textual
transcriptions of written posts, spoken turns, bodily activity
of avatars, and onscreen activity by users that occur in CMC
data, and in Section <ptr/> we gave recommendations on how to
encode graphics or other media data within posts with
<att>modality</att> set to <val>written</val>. When two or more
of these features occur in a CMC interaction, we can speak of
<hi rend="italic">multimodal</hi> CMC.</p>
<p>As explained in Section <ptr target="#CMCUnits"/> the elements <gi>post</gi>, <gi>u</gi>,
<gi>kinesic</gi>, and <gi>incident</gi> are available to to encode textual transcriptions
of written posts, spoken turns, bodily activity of avatars, and onscreen activity by users
that occur in CMC data, and in Section <ptr/> we gave recommendations on how to encode
graphics or other media data within posts with <att>modality</att> set to
<val>written</val>. When two or more of these features occur in a CMC interaction, we can
speak of <hi rend="italic">multimodal</hi> CMC.</p>

<p>Some basic multimodality is available in private chat such as WhatsApp, where spoken and
written posts and media posts containing images or video clips, can alternate. The following
Expand Down

0 comments on commit 5a4e803

Please sign in to comment.