Skip to main content

A Real-Life Example

This page is under construction

What follows is an account of a quarrel about the definition of 'identity', that took place in a meeting of an EU project. The account shows one of the simpler interventions that can lead to the resolve such discussions. "In the wild", things are usually not this simple, but a story like the one below does provide an idea of what we're after.

1. Real-Life Observations​

The problems that a lack of linguistic unity can pose are widely recognized. The typical reflex to overcome them is to work to create a glossary, i.e. an (alphabetically sorted) list of terms, each of which is followed by a description, which may include, or be, a definition. Such glossaries are the result of discussions in which it is decided which terms are to be included, and what their definitions (descriptions) should be. Discussions about the definition of terms are not limited to the creation of glossaries. They can pop up in arbitrary situations where different people work together.

It seems as if the mechanism is that the (most prominent) participants in the discussion each have a (strong) idea of what the definition of the term is, that these definitions differ in some significant way, and the participants then either try to convince the others of their truth, or they try to find a definition that they all agree on. It may also happen that the group leader at some point ends the discussion, with or without stating what the definition is going to be from that point forward.

Each of these ways leads the point where not only a definition is established, but also where multiple, if not all participants are left with a feeling of dissatisfaction, if not frustration, that it still isn't right. This means that participants are unlikely to commit to using that definition, e.g. as they author documents, which implies that the quality of the document suffers as readers cannot rely on the term meaning what its definition says.

Also, these feelings will, after some time, trigger participants to rekindle the discussion. But alas, as the method used to come to resolve the issue hasn't changed, it must be expected that the result, even if the definition itself is changed, remains the same: people remain unsatisfied, they are not committed to use the definition, and the quality of the work continues to suffer.

2. Changing the discussion process​

The crux to resolving such issues is to change one's perspective from focusing on the term (and a definition to go with it), to focussing on the meanings that cause the dispute.1

In the EU project, three participants, which we will refer to as P1, P2 and P3, had a stiff argument about the definition of the term 'identity', which unfolded pretty much as described above. We then intervened by asking if they would agree to an experiment to resolve the issue. They gave their consent, so we started the following process.

2.1. Get to understand the differences​

We first asked P1 to phrase the criterion that he used to distinguish between what is, and what is not, an identity. He said something like "Well, its things like passports, driving licenses, things like that, you know... ". Some people objected, saying that this is not really a criterion. We accepted the comment, and deferred the discussion to a later time.2

Going back to what P1 said, we took up a cup, asked all participants to apply the criterion as stated to the cup, and tell us what they would conclude after having applied the criterion. There are three possibilities: the cup satisfies the criterion, it does not, or it cannot be determined. There was agreement that the cup wasn't something like a passport, driving license, or things like that. We took out an ID-card, asked the same question, and again we had agreement: yes, this is something like a passport, driving license, or things like that. We then stated that this shows that all participants can make the same distinction as P1 does when they use this criterion, and hence they understand what P1 means when he talks about 'identity'.

We then asked P2 to tell us his criterion for distinguishing identities from non-identities. He told us: "It's not the passports and such, but its what's in there: name, address, date of birth and such." Again we pointed to the cup, and there was consensus that the cup didn't satisfy the criterion. We then pointed to the ID-card, and there was consensus that the data on the card did satisfy the criterion. We concluded that we all understood what P2 meant when he talked about 'identity'. It was also clear that this meaning differs from what P1 used, and that's quite ok.

Asking P3 for his criterion, he said: "its not all the data, but just the id-number, like a BSN (for the Netherlands), a Social Security Number, or such". It was obvious that the cup did not satisfy the criterion, and the government identifier on the ID-card did. Then, it was noticed that passports, driving licenses and the like have their own serial numbers, that also satisfy this requirement. So we asked P3 whether or not this matched his idea of identities, which he denied. Then all joined in to reformulate the criterion that would allow all of us to distinguish between what P3 did, and did not call an 'identity', as follows: "an alphanumeric string that is used to identify a natural person".

2.2. Check relevance, and assign names​

There was agreement that the criteria distinguished between different things: a passport is not its data content, which again is distinct from the value in a single data field. We then asked which of these distinctions were relevant, and which were not, within the context of our project. After all, we do not want to occupy ourselves with distinctions that are irrelevant to our cooperation. There was agreement: they were all relevant.

Since we now had three meanings (criteria) that were considered relevant for our work, it is beneficial to establish terms to refer to each of these three. From a theoretical point of view, it doesn't really matter what the terms are - you could decide to call them 'ID1', 'ID2' and 'ID3', as long as the members in the project would agree on them, and commit to using them.

In practice, however, it is beneficial to use terms that correspond with terms that are already in practical use. But it comes with the risk that people will take these terms in the meaning that they are used to, rather than in the meaning as specified by the criteria. That's why it is necessary that all participants commit to using this 'new' terminology in the ways that they have specified, not just in their speech, but also - and particularly - when authoring documents.

In the discussion that followed, the term:

  • 'identity document' was assigned to P1's criterion, so passports, driving licenses etc. would be called an identity document;
  • 'identifier' was assigned to P3's criterion, so every alphanumeric string that is used to identify a natural person, such as a BSN or a social security number, would from henceforth be called an identifier.
  • 'identity' was assigned to P2's criterion, which was refined to "a set of data (attributes, claims) about a single person".

3. Reflecting on the process​

Reflecting on what has happened here, there are several observations to be made.

3.1. A basic activity​

One of the basic activities is that a single person (the 'protagonist' of the activity) gets to explain his understanding of the term, but not in a freewheeling fashion. He is asked to provide a specific criterion that all can later evaluate, and that distinguishes between what the person understands to be instances or non-instances of the term. This isn't easy. There is a simple test to demonstrate that what people know best is difficult for them to express in words.3 So we do not expect that person to come up with a first-time-right criterion; we will work with what the person provides.

We may need to explain to the others that this process is not about judging whether or not this person is wrong, nor if he is right, nor if what he comes up with is a 'real' criterion. The only two judgements that participants are allowed to make are:

  • whether or not they can evaluate the criterion that this person has provided, i.e. wether or not they will be able to answer with 'yes' or 'no' to the question of whether something that is identified in a use-case or example, satisfies the criterion, or not.
  • whether or not something that is identified in a use-case or example, satisfies the criterion.

We require the other participants to listen, and try to understand what the person is trying to express. This, too, isn't easy - not for everybody. So we also do not expect them to be able to do this, we do attempt to make them focus on

  • making the two judgements as explained in the previous paragraph, and
  • thinking about use-cases or examples that may provide a deeper insight into the actual distinction that the person is trying to make.

By confining the tasks of each participant to the ones mentioned above, the discussion doesn't go all over the place.4 Also, because the tasks are aligned to produce a well-defined result (i.e. a criterion that all use in the same way, and produces the same judgements across the participants when applied to use-cases/examples), we see that the fighting is dramatically reduced, and often completely stops. Instead, participants feel more respected, and heard.

This is further stimulated by having participants help the protagonist, not just by coming up with examples to 'test' the criterion, but also by proposing alternative formulations for the criterion that the protagonist is trying to express - formulations that might be easier to evaluate, or that make a more precise distinction. 'Helping the protagonist' means that the protagonist is the one that has the ultimate say in what works, and what does not. After all, it's his 'knowledge' that is being made explicit.

3.2. Bringing it all together​

The basic activity is run multiple times, giving every participant the chance to be the protagonist. This results in a multitude of criteria that all have a sufficiently similar understanding of, and of which the relevance is acknowledged by all.

Note that the process has a side effect that may be called 'team building'. When participants diligently work together to really understand each other, they will feel heard, and respected in their thinking, even if it differs from that of others. Making the differences more explicit and treating them as the real contributions that they are also helps participants to become more friendly towards each other.

This helps in the next step, which is to come up and agree on the terms that are to be associated with the various criteria (that make the relevant distinctions). This discussion differs from the one that was stopped (and replaced by the process) in the sense that it is no longer a matter of what the 'right' or 'correct' meaning of a term is, or should be. Rather, it is a discussion about what the most appropriate word or phrase is to refer to an agreed-upon criterion. Participants would feel that it is the criterion that really matters, and the chosen term is not unimportant, but secondary.

Time that is spent to choose terms wisely isn't wasted. After all, terms will not only be used within the group, but they will also need to be interpreted by people outside the group, e.g. those that read documents, or hear talks that originate from (one or more members of) the group. If a term is chosen that doesn't ring bells with an external audience, it means that it has to be explained all the time (which could be beneficial). If a term is chosen that is known to be used in various meanings, crafting documents or presentations need to make sure that the meaning as is intended, is actually conveyed to the audiences.

Then, participants need to commit to using this 'new' terminology in the ways that they have specified, i.e. such that it makes the distinction as intended by the associated criterion. Participants should use it as such not just in their speech, but also - and particularly - when authoring documents. And when their utterances are intended for external audiences, they should make sure such audiences are pointed to the definitions (the criteria) that specify the intended meaning.

3.3. Consolidation​

The process of coming up with criteria and assigning words/phrases to them can be concluded by

  1. writing a glossary, i.e. an (alphabetically sorted) list of the terms and their definitions (criteria). However, such a glossary differs from a glossary that is created by the process described in section 1. A glossary that results from the process we advocate consists of terms that are all associated with a criterion that users can evaluate in different situations (that are relevant for the purposes of the scope of which the glossary is part) and thus determine what is, or is not, an instance of the term. This is a property that is not very common to see in other glossaries.

  2. further documenting each term, e.g. by providing some additional descriptions that are intended to help non-participants to get to the same level of understanding of the term as the participants have acquired by discussing it, providing use-cases to test the criterion, etc. Such descriptions might contain a further elaboration of the term, a few examples (e.g., from the use-cases), a specification of the purposes(s) for which the term is particularly useful, etc.

  3. documenting the coherence between the terms, e.g. as one or more mental models. They serve similar purposes as the additional descriptions of terms, but rather than drilling down on the meaning of a single term, they focus on the relations that terms have between each other, and constraints that apply. Patterns thus provide additional semantics that are difficult to capture in their entirety in the definition and/or further description of single terms.

3.4. Tools for consolidation​

Many tools exist to document various things, particularly in the context of IT, where the use of e.g. GIT repositories is common. Tools such as Docusaurus, GitHub Pages, Jekyll, and many more. These tools can be used to write a glossary, and document terms and mental models.

Also, a toolbox called TEv2 is being constructed that aims to provide support for this, e.g. by automatically generating glossaries, and providing enabling authors to refer to specific definitions and descriptions of terms that they use, making it easier for readers of that documentation to know the meaning thereof that the author intended.

Notes​

  1. See also: R. Joosten, 'On Terminology and the Resolution of Related Issues', TNO, Technical Report R11793, DOI: 10.13140/RG.2.2.17546.18888.↩
  2. The discussion about what is (not) a 'criterion' is a different topic. We don't want to discuss two definitions at the same time.↩
  3. People that are native speakers of a language, can usually demonstrate their knowledge of the specifications of that language (its grammar), by showing that they can (1) determine whether or not a sentence in that language is grammatically correct, (2) correct the sentences that were incorrect and (3) come up with (say) 5 alternatives for correcting an incorrect sentence, and picking out 'the most beautiful', or 'the best' one. However, when asked whether they could write down the specifications such that the result is consistent, coherent, complete, cogent, congruent - and other c-words that refer to characteristics that specifications are typically expected to have, they will say 'no'. Even worse: if they are presented a booklet that is about the grammar of their mother's tongue, and are asked to determine whether or not it specifies that grammar in a consistent, coherent, complete, etc. way, they will also decline.↩
  4. Confining what participants can do helps to keep the activity focused. This is particularly important if the process is under time constraints (which most of the time is the case). There are various techniques that contribute to keeping the focus, two of which are already mentioned. Another technique is to refuse to talk about/process any use-case/example that is irrelevant for the objectives that the participants, as a group, try to realize.↩