Clean Language in Research Interviews

Gathering data without leading
Print Friendly, PDF & Email

Publisged in Rapport, Issue 40, August 2014

The purpose of this article is to outline applications of Clean Language to research interviewing. While Clean Language is a distinct approach from NLP, there is an affinity between them in their emphasis on the relationship between language and people’s inner worlds. Hence NLP practitioners are likely to find Clean Language accessible and of potential use as an approach to data gathering through interviewing.

Research interviewing

Interviewing is probably the most commonly used approach to data-gathering in qualitative research,(*1) being employed across disciplines that include anthropology, sociology, psychology and organisational studies. A typical definition of an interview is ‘a conversation with a purpose’(*2) and according to King,(*1) ‘the goal of any qualitative research interview is…to see the research topic from the perspective of the interviewee, and to understand how and why they come to have this particular perspective’.

Like any other data collection method, interviewing may be more or less useful depending on its fit with the research aims and objectives. It offers advantages such as flexibility (it can be used almost anywhere), the capacity to explore meaning in depth, and a more personal engagement than is possible with (for example) questionnaires. On the other hand, interviewing is time consuming and resource intensive, especially when it comes to transcribing and analysing data.

Investigating the potential of Clean Language

Building on prior interests in ways in which NLP’s approach to questions can be applied to qualitative research interviewing, specifically phenomenology (as described in a previous article in this series), I collaborated with a group of practitioners (*3) on a project that investigated the potential of Clean Language to enhance the validity of research interviewing. For the full details of that project and its findings, see Tosey et al.(*4)

Clean Language is a practice that aims to facilitate exploration of a person’s inner
world through their own, naturally occurring metaphors. Its origins lie in the work of counselling psychologist David Grove in the 1980s,(*5,6) who discovered that by focusing on a client’s naturally occurring metaphors and enabling that client to become immersed in their inner symbolic world, spontaneous change could occur. Grove called this approach ‘Clean Language’ (*5) because of its intention to keep the practitioner’s language as ‘clean’ as possible, in the sense of being free from the practitioner’s own metaphors. This helps to avoid (in NLP terminology) ‘leading’ the client.

Grove identified a basic set of Clean Language questions (*7 for example) which he considered to be the most minimal in terms of avoiding the introduction of the practitioner’s constructs and metaphors. Two examples of clean questions are, ‘is there anything else about X?’ and ‘what kind of X is that?’ where X stands for the interviewee’s exact words.

Introduced content

According to the work of Lakoff and Johnson, who defined the essence of metaphor as ‘understanding and experiencing one kind of thing in terms of another’,(*8) it is both easy and common for an interviewer to introduce their own metaphors inadvertently. There are two main stages at which this can happen, both of which represent threats to the validity of interview-based research.

First, the interviewer can include their own content within a question, thereby bringing the interviewer’s constructs and metaphors to the interviewee’s attention and potentially influencing their responses. Second, the interviewer can introduce their own metaphors when analysing and writing up the research (for an example, see Tosey et al.).(*4) These issues are compounded by the fact that the full detail of questions asked and responses received is rarely disclosed in published research, so that it can be difficult for the reader to make their own assessment of the relationship between the questions asked, the data and the findings.

The research project

Our research project interviewed a small sample of six managers relating to the way they experienced work-life balance (WLB). This involved an adjustment to the way Clean Language is used, in that it is normally led by the participant’s outcome (such as a desire for change, or a goal of behavioural improvement). In this project the reason for conducting the interview was to meet a research aim rather than to meet the participant’s outcome; personal change was deliberately not pursued and was not part of the contract with the participant. Nevertheless, it was interesting that in a follow-up interview, several participants reported that they had decided to make changes based on the awareness they gained through the first interview. From an ethical perspective it is also relevant to note that in a subsequent follow up enquiry, all participants reported that they had found the interview beneficial, and no participant identified any disbenefits. Interviewing is acknowledged to be a powerful process and the project fully observed the institution’s research ethics guidelines.

Considerable time and care was taken over deciding how to start the interviews, and especially how to introduce the topic of WLB given that ‘balance’ is itself a metaphor. We took the decision to use the term WLB at the start of the interview in order to introduce the topic in a way that the interviewer could easily understand, but then to focus firmly on the interviewer’s own words.

All the interviews were recorded, transcribed then reviewed. First, all the questions used by the interviewer were looked at it in terms of the extent to which they were ‘clean’, including whether they were confined to using the interviewee’s own language. In this project the interviewer did remarkably well, with more than 95 per cent
of questions being considered ‘clean’. Eleven basic Clean Language questions accounted for 85 per cent of the interview questions, which underlines how useful those basic questions can be. The analyses were then scrutinised by the team in order to confirm that they represented the interviewee’s own metaphor landscape as faithfully as possible according to the data. An advantage of Clean Language is that, rather than relying on our individual interpretations of the material, we were guided by a common set of principles.

The analysis yielded individual metaphor landscapes for each manager. While there is not space to describe these here, it was notable that no participant used a metaphor of ‘balance’, although several of the metaphors entailed notions of ‘balancing’ (e.g. juggling). This is interesting because it means that many individuals might not relate directly to the notion of work-life ‘balance’ that is in common usage.


For people familiar with NLP, one interesting difference between Clean Language interviewing and a standard qualitative interview concerns rapport. It is typically assumed that rapport, reflected for example in eye contact and congruent non-verbal behaviour, is helpful in an interaction between people. While establishing a context of rapport at the outset would still be important in a Clean Language interview in order to enhance trust and openness, within the interview itself the interviewee’s attention needs to be on their own inner world or metaphor landscape. The interviewer could interrupt that inward attention through using techniques such as eye contact and mirroring, which would therefore be counter-productive.

In conclusion

Although Clean Language is especially suitable when eliciting ‘metaphor landscapes’, clean questions can be incorporated into any interview. They can therefore contribute to rigour and quality in the use of qualitative interviews by minimising the inadvertent introduction of the interviewer’s constructs. More extensive use of Clean Language – for example in order to elicit metaphors or to explore the participant’s inner world – is likely to need the interviewer to have undergone training in Clean Language (for further detail see Tosey et al.).(*4)

In our project the interviewer had considerable experience of Clean Language and the project team included acknowledged experts in the field. In conclusion, the WLB project illustrates the potential for Clean Language as a method of metaphor elicitation specifically, and as a way of enhancing the quality of research interviewing more generally. It provides not only a distinctive approach to questioning (see also Tosey)(*9) but also a systematic and rigorous way of exploring and maintaining fidelity to a person’s own inner world, thereby contributing to important criteria for validity in qualitative research such as authenticity and trustworthiness.(*10) These represent enhancements that can be transferred from the world of practice to that of research.


(*1)  N. King, ‘Using interviews in qualitative research’. In C. Cassell and G. Symon (eds), Essential Guide to Qualitative Methods in Organizational Research, pp. 11–22 (2004).

(*2)  B.L. Berg, Qualitative Research Methods for the Social Sciences.

(*3)  Wendy Sullivan, Margaret Meyer and Rupert Meese of the Clean Change Company, and James Lawley of the Developing Company. Sarah Nixon, Liverpool John Moores University, acted as advisor to the project.

(*4)  P. Tosey, J. Lawley and R. Meese, ‘Eliciting metaphor through Clean Language: an innovation in qualitative research’. British Journal of Management, Vol. 25, 629–646 (2014)

(*5)  D.J. Grove and B.I Panzer, Resolving Traumatic Memories: Metaphors and Symbols in Psychotherapy.

(*6) J. Lawley and P. Tompkins, Metaphors in Mind: transformation through symbolic modelling.

(*7) W. Sullivan and J. Rees, Clean Language: revealing metaphors and opening minds.

(*8) G. Lakoff and M. Johnson, Metaphors We Live By.

(*9) P. Tosey, ‘And what kind of question is that? Thinking about the function of questions in qualitative interviewing’. In Handbook of Research Methods on Human Resource Development.
Editors, Saunders, M. N. K. & Tosey, P. (2015).

(*10) Y. Lincoln and E. Guba, Naturalistic Inquiry.

Print Friendly, PDF & Email
body * { color: inherit !important; }