The concept of function creep

Abstract

Function creep – the expansion of a system or technology beyond its original purposes – is a well-known phenomenon. Correction: it is a well-referenced phenomenon. Yearly, hundreds of publications use the term to criticise developments in technology regulation and data governance, but surprisingly, no-one has ever written a paper about the concept itself. This paper fills that gap in the literature, by analysing and defining ‘function creep’. This creates conceptual clarity that can help structure future debates and address function creep concerns. After analysing the term ‘function creep’ itself, I discuss concepts that share family resemblances, including other ‘creep’ concepts and many theoretical notions from STS, economics, sociology, public policy, law, and discourse theory. Function creep can be situated in the nexus of reverse adaptation and self-augmentation of technology, incrementalism and disruption in policy and innovation, policy spillovers, ratchet effects, transformative use, and slippery slope argumentation. Based on this, I define function creep as an imperceptibly transformative and therewith contestable change in a data-processing system’s proper activity. Argumentation theory illuminates how the pejorative ‘function creep’ functions in debates: it makes visible that what looks like linear change is actually non-linear, and simultaneously calls for a much-needed debate about this qualitative change.

Full text

Full Terms & Conditions of access and use can be found at
https://www.tandfonline.com/action/journalInformation?journalCode=rlit20
Law, Innovation and Technology
ISSN: (Print) (Online) Journal homepage: https://www.tandfonline.com/loi/rlit20
The concept of function creep
Bert-Jaap Koops
To cite this article: Bert-Jaap Koops (2021): The concept of function creep, Law, Innovation and
Technology, DOI: 10.1080/17579961.2021.1898299
To link to this article: https://doi.org/10.1080/17579961.2021.1898299
© 2021 The Author(s). Published by Informa
UK Limited, trading as Taylor & Francis
Group
Published online: 16 Mar 2021.
Submit your article to this journal
View related articles
View Crossmark data

The concept of function creep
Bert-Jaap Koops
TILT, Tilburg University, The Netherlands
ABSTRACT
Function creep –the expansion of a system or technology beyond its original
purposes –is a well-known phenomenon. Correction: it is a well-referenced
phenomenon. Yearly, hundreds of publications use the term to criticise
developments in technology regulation and data governance, but surprisingly,
no-one has ever written a paper about the concept itself. This paper fills that
gap in the literature, by analysing and defining ‘function creep’.Thiscreates
conceptual clarity that can help structure future debates and address function
creep concerns. After analysing the term ‘function creep’itself, I discuss
concepts that share family resemblances, including other ‘creep’concepts and
many theoretical notions from STS, economics, sociology, public policy, law, and
discourse theory. Function creep can be situated in the nexus of reverse
adaptation and self-augmentation of technology, incrementalism and disruption
in policy and innovation, policy spillovers, ratchet effects, transformative use,
and slippery slope argumentation. Based on this, I define function creep as an
imperceptibly transformative and therewith contestable change in a data-
processing system’sproperactivity. Argumentation theory illuminates how the
pejorative ‘function creep’functions in debates: it makes visible that what looks
like linear change is actually non-linear, and simultaneously calls for a much-
needed debate about this qualitative change.
ARTICLE HISTORY Received 8 October 2019; Accepted 29 January 2020
KEYWORDS Function creep; creep concepts; databases; social systems; argumentation theory;
conceptualisation
1. Introduction
Function creep is a phenomenon familiar to most scholars in the fields of
Science & Technology Studies, law and technology, and Surveillance
Studies, and to many other scholars interested in how technologies and
information systems are used and regulated in society. It is so familiar that
authors typically use the term without feeling a need to define or explain
it. At most, they briefly describe the phenomenon in a few words, assuming
© 2021 The Author(s). Published by Informa UK Limited, trading as Taylor & Francis Group
This is an Open Access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDer-
ivatives License (http://creativecommons.org/licenses/by-nc-nd/4.0/), which permits non-commercial re-use, distri-
bution, and reproduction in any medium, provided the original work is properly cited, and is not altered,
transformed, or built upon in any way.
CONTACT Bert-Jaap Koops e.j.koops@uvt.nl
LAW, INNOVATION AND TECHNOLOGY
https://doi.org/10.1080/17579961.2021.1898299

that readers will know what they are referring to. We all know it has some-
thing to do with a gradual expansion of the functionality of some system or
technology beyond what it was originally created for.
But why exactly is ‘gradual function expansion’a concern, and why do
authors label this phenomenon –pejoratively –‘function creep’? The wide-
spread use of the term indicates a prevalent concern with something going
wrong, or at least not quite right, when a system
1
acquires new uses. Appar-
ently, function creep is something to be addressed, and therefore, an impor-
tant phenomenon in our effort to understand and regulate technology.
‘Creep’has many different connotations (e.g. slowness, invisibility, stealth,
uncanniness), and the literature is not at all clear or coherent on what
exactly is wrong with function creep and what should be done about it.
Wherein exactly lies the ‘creepiness’of function creep? If we do not under-
stand the core of function creep, it will be hard to find suitable responses to
address the concern that many authors voice when calling something ‘func-
tion creep’.
Surprisingly, the concept of function creep has never been analysed, at
least not in any real depth. No literature is available on defining ‘function
creep’or explaining why it causes concern. Even Wikipedia (at the time of
writing) conspicuously lacks a page on function creep. As far as I have
been able to establish, the most elaborate analysis of function creep is a
1500-word conceptual analysis by Johanne Yttri Dahl and Ann Rudinow
Sætnan in their paper on governing DNA databases.
2
Their analysis is excel-
lent, but short and far from comprehensive. For a concept so frequently used
in social-science literature, function creep merits an analysis of its own –not
focused on a particular system, but on the meaning and implications of the
concept generally.
In this paper, I analyse the concept of ‘function creep’in order to develop
adefinition of the term. Having a precise definition is important, because
conceptual clarity will help structure future debates, about function creep
in general and specific instantiations of it. Moreover, analysing the
concept may yield insight into how and why function creep occurs, which
helps us understand how we can deal with it. Therewith, this paper provides
groundwork on which future scholarship can build to address the challenges
of function creep.
To this effect, in section 2, I analyse what ‘function creep’refers to, based
on a semiotic analysis of its constituent parts (‘function’and ‘creep’) and of
the compound (‘function creep’), including a discussion of its role in
1
The thing of which the function creeps can vary –it can be a technology, technological application,
system, database, project, programme, or even a law (associated with some system or database).
For convenience’sake, in this paper, I use ‘system’to refer to the thing that displays function creep.
2
JY Dahl and AR Saetnan, ‘“It All Happened So Slowly”–On Controlling Function Creep in Forensic DNA
Databases’(2009) 37 International Journal of Law, Crime and Justice 83, 84–7.
2B.-J. KOOPS

discourse (how and why it is used in academic and policy debates). In section
3, I discuss concepts from many different fields that share family resem-
blances with ‘function creep’. This helps to understand the linguistic niche
that the term ‘function creep’occupies among conceptually related theoreti-
cal notions; it also provides important insights into possible mechanisms
underlying function creep. These insights are used in section 4 to develop
adefinition of function creep. Section 5 offers a brief conclusion.
2. Definitions and meanings
The first path to understanding the concept of function creep is to have a
close look at definitions given in authoritative dictionaries and in the litera-
ture, both of the two constitutive elements of the expression –‘function’and
‘creep’–and of the whole.
2.1. Function
According to the Oxford English Dictionary, ‘function’(noun) has six poss-
ible meanings. The primary meaning in the etymological sense is: ‘The action
of performing; discharge or performance of (something)’.
3
Thus, the func-
tion creep of X has something to do with the performance of X or with
the way in which X functions. But the function of X is not merely restricted
to the actual act of performing. A more specific meaning of ‘function’that is
the most relevant in ‘function creep’is:
The special kind of activity proper to anything; the mode of action by which it
fulfils its purpose (…)c. of things in general.
4
Thus, the function of X indicates not merely the activity of X in practice,
but the regular or proper activity of Xin relation to its purpose. The func-
tion of a hammer is to hit a nail: it fulfils its purpose when used for hitting a
nail. It is debatable whether the function is broader, say hitting the skull of
someone you want to kill; the hammer can certainly be used for such a
purpose, but many would not say that this is its purpose, in the sense of
hitting people being proper to hammer use. Somehow, the function of X
is considered something ‘intrinsic’to X,inthesenseofXbeing meant
to perform certain actions. This raises an important question for under-
standing function creep: who determines a system’s‘proper’activity or
‘intrinsic’function? Is it the one who commissions the application, the
developer, or the user? From the field of Science & Technology Studies,
3
JA Simpson and ESC Weiner (eds), The Oxford English Dictionary (Clarendon Press, 2nd edn 1989).
4
(n 3). The other meanings of ‘function’are: an activity, action in general; the kind of action proper to a
person as belonging to a particular class, hence the office itself, an employment, profession, calling,
trade; a ceremony; a type of mathematical expression.
LAW, INNOVATION AND TECHNOLOGY 3

we know that there is no obvious answer to this question: it depends on the
circumstances whether ‘the’function of an application is determined more
by its design specifications, by the functionalities the designer –intention-
ally or unwittingly –builds in, or by how users actually use the
application.
5
2.2. Creep
The ‘creep’in ‘function creep’is a noun that refers to the ‘action of creeping
(lit. and fig.)’.
6
The verb ‘to creep’,asdefined in the Oxford English Diction-
ary, has several meanings. The ‘creep’in ‘function creep’is more equivocal
than ‘function’: it is not quite clear which of its many meanings may be
evoked in ‘function creep’. The first three meanings are related and certainly
relevant:
1. To move with the body prone and close to the ground, as a short-legged
reptile, an insect, a quadruped moving stealthily, a human being on hands
and feet, or in a crouching posture (…).
2. a. To move softly, cautiously, timorously, or slowly; to move quietly and
stealthily so as to elude observation (…).
b. Of things: To move slowly.
c. To introduce gradually; slowly to increase (an amount of light, volume of
music, etc.).
3. fig. (of persons and things). a. To advance or come on slowly, stealthily, or by
imperceptible degrees; to insinuate oneself into; to come in or up unobserved
(…).
b. To move timidly or diffidently; to proceed humbly, abjectly, or servilely, to
cringe; to move on a low level, without soaring or aspiring.
7
The common element in these meanings is a particular type of movement
that is (1) slow, gradual and/or (2) stealthy, eluding (or intended to elude)
observation. Both elements seem relevant for function creep.
The first element can indicate that a function is changing slowly,
8
but it
can also be important that the process of function creep happens gradually,
step by step.
The second element seems to have gained importance over the past
decades in the meaning of ‘creep’. The current Oxford English Dictionary,
5
Cf N Oudshoorn and T Pinch, How Users Matter: The Co-construction of Users and Technologies (MIT Press,
2003); B Latour, ‘Where are the Missing Masses? The Sociology of a Few Mundane Artefacts’in WE
Bijker and J Law (eds), Shaping Technology-Building Society. Studies in Sociotechnical Change (MIT
Press, 1992) 225.
6
(n 3), under ‘creep’(substantive).
7
(n 3), under ‘creep’(verb).
8
Cf the title of Dahl and Saetnan’s article (n 2).
4B.-J. KOOPS

as compared to its predecessor (the 1973 Shorter OED), added the words ‘so
as to elude observation’in the second meaning, and in the third meaning
changed ‘advance (…) by degrees’into ‘advance (…)byimperceptible
degrees’. This element is certainly important for the concept of function
creep, indicating that the change in X’s function happens imperceptibly.
This could be intentional (‘so as to elude observation’) but it could also
happen factually (‘to advance by imperceptible degrees’). It can have a
neutral connotation (‘to come in unobserved’) but it will often have a some-
what negative association (‘to insinuate oneself’;‘stealthily’, which derives
from stealing, indicating a furtive action, ‘secretly, clandestinely’
9
).
A negative association of ‘function creep’could also be related to the fact
that ‘creep’is a typical movement of reptiles or insects and that the move-
ment thus has come to be associated with ‘creepy’things. The sixth
meaning of ‘creep’(verb) is therefore also relevant:
6. Of the skin or flesh, less usually of the person himself. To have a sensation as
of things creeping over the skin; to be affected with a nervous shrinking or
shiver (as a result of fear, horror, or repugnance).
10
A creeping movement might therefore also be ‘creepy’: tending to produce
the sensations of ‘a creeping of the flesh, or chill shuddering feeling,
caused by horror or repugnance’.
11
Possibly, people could perceive a
change in the function of X as repugnant or horrible, as it gives them the
‘creeps’, i.e. a sensation as of creeping things on one’s body.
12
However,
this ‘creepy’feeling is not necessarily evoked by ‘function creep’. The nega-
tive association might also consist in the sneakiness of the process, triggering
associations of an imperceptible snake-like movement, rather than associ-
ations of the creepiness of an insect-like movement on the skin.
Possibly, for some people, the ‘creep’in function creep might also raise
connotations of personified creepiness, as in the colloquial meaning of a
‘creep’as a ‘creeping fellow; a sneak’, or the slang meaning of a ‘despicable,
worthless, stupid, or tiresome person’.
13
Using the term ‘function creep’to
warn against a certain risk might in that sense also (largely implicitly and
subconsciously) capitalise on people’s ingrained fear of encountering a
‘creep’in places where they feel vulnerable.
9
(n 3) under ‘stealth’.
9
(n 3) under ‘stealth’.
10
(n 3). The other meanings listed in the dictionary do not add much to our understanding of function
creep: 4. movement of plants along the ground or surfaces; 5. creep along or over; to rob (stealthily);
7. to drag in deep water with a creeper; 8. movement of metal rails etc. under pressure; 9. to suffer a
‘creep’in coal mining; 10. to move imperceptibly en masse of soil. 11. of a rubber tyre; 12. to slip or
slide backwards of a belt or rope.
11
(n 3) under ‘creepy’.
12
(n 3) under ‘creep’(substantive).
13
Ibid.
LAW, INNOVATION AND TECHNOLOGY 5

In sum, the element ‘creep’in ‘function creep’is a complex term, which
primarily triggers associations of slowness and/or imperceptibility of move-
ment, but which at the same time is strongly associated with negative conno-
tations of stealth or furtivity, and possibly also of feelings of fear or
repugnance.
2.3. Function creep
2.3.1. Origins and prevalence of the term
The term ‘function creep’is nowadays frequently used in academic literature
and, to a lesser extent, in popular discourse. A search on the term ‘function
creep’in English-language newspapers yields over 600 results.
14
The term
seems to have gained currency in the second half of the 1980s in Australia.
Newspapers first mention the term in October 1988 in relation to a tax file
number plan,
15
and Australian newspapers soon pick up the term in relation
to other developments qualified as surveillance measures.
16
The term then
hops over to North America, particularly to Canada in discussions of
health cards.
17
Use of the term is largely restricted to Australia and
Canada until the late 1990s, when the United States and British newspapers
start using it too.
18
More recently, the term is occasionally used in other
countries’news.
19
The term, however, remains used considerably more fre-
quently in Australian newspapers, followed by Canadian newspapers, than in
other countries’news.
That the concept of function creep originated in 1980s Australia is
confirmed by a search in academic sources. Google Scholar yields over
1600 results,
20
the earliest of which is Roger Clarke’s discussion of the
14
Based on a search in Lexis Nexis Academic in all English-language newspapers, September 13, 2018.
The search yielded 668 hits, but these included four academic journals wrongly classified by Lexis Nexis
as news outlets, around 50 double entries, and a few articles on ‘creep’in a material-technical sense
(e.g., about patenting a ‘creep-suppression function’), so the number of relevant hits is little over 600.
An earlier search, on August 29, 2012, yielded 545 results, suggesting that there is no particularly
strong rise in usage recently. However, the new search has 145 entries since August 29, 2012, so
the two sets are not completely comparable.
15
‘Report May Kill Tax File Number Plan’Herald (20 October 1988).
16
See, e.g., ‘Big Brother and the Ultimate Smart Card’Australian Financial Review (15 May 1989); ‘Big
Brother is Watching, and Waiting’Sydney Morning Herald (5 September 1990); ‘The Nation Decides.
New ID Card “By Stealth”’ Sunday Mail (South Australia) (28 February 1993).
17
See, e.g., ‘Privacy Concerns are Raised Over Health Card Plan’The Atlanta Journal and Constitution (24
September 1993); ‘Pharmanet is Scary Medicine’The Vancouver Sun (British Columbia) (31 March 1994);
‘“Smart”Cards Under Attack: Use Will Expand Beyond Original Intent: Critics’Hamilton Spectator
(Ontario) (6 November 1995).
18
See, e.g., ‘Plan Would Let Private Labs Do DNA Tests’USA TODAY (8 June 1998); ‘It’s Too Late, Big
Brother is Here and He’s All Eyes’The Scotsman (14 December 1998); ‘The People’s Plug: What can
HMG.org offer us? Track?’The Guardian (25 March 1999).
19
E.g., ‘UID Bill Skips Vital Privacy Issues’Indian Express (26 September 2010); ‘Biometric Smart ID Cards –
Dumb Idea’Africa News (9 July 2014).
20
https://scholar.google.com, search query ‘“function creep”-physics -materials -fatigue -elastic -delta’,
March 14, 2019. The exclusion of physics etc. ensures that most references to ‘function creep’in the
material, not social, sense are eliminated.
6B.-J. KOOPS

Australian tax file number system.
21
This system ‘has exhibited the charac-
teristic popularly referred to as “function creep”, whereby additional uses
accumulate, and change the purpose of the scheme.’
22
Apparently, by 1991
the term was already in popular use down under. The term continues to
be used by Clarke over the next years, and is picked up by other scholars
and privacy experts by 1994.
23
During the 1990s, the term is used fairly
seldom in academic publications, but its frequency rises steadily and fairly
steeply in the 2000s, reaching a peak around 2010 and stabilising there at
well over a 100 publications per year.
24
Thus, both in news reports and aca-
demic publications, the term got used increasingly throughout the 1990s and
early 2000s, but its usage over the past decade is relatively stable with a sig-
nificant but not extremely high prevalence.
2.3.2. Definitions of the term
The term ‘function creep’is not (yet) defined in most authoritative English-
language dictionaries or encyclopaedias. Apparently, the concept is too
young for this or has not been used frequently enough in the corpora used
by most dictionaries. It is defined, however, in Collins Dictionary as ‘the
gradual widening of the use of a technology or system beyond the purpose
for which it was originally intended, esp when this leads to potential invasion
of privacy’.
25
Surprisingly, the term still does not have an entry in Wikipedia at the time
of writing (August 2019). Wikipedia automatically redirects the search entry
‘function creep’to ‘scope creep’(see section 3.1). Although related, it is more
limited in meaning, as function creep is used not only for projects, but also
for systems, technology applications, laws, and databases, usually involving
some form of (personal) data processing.
In academic literature, there is no commonly accepted definition of func-
tion creep. Most authors who use the term define or describe it in their own
21
R Clarke, ‘Tax File Number Scheme: A Case Study of Political Assurances and Function Creep’(1991) 7(4)
Policy.
22
Clarke (n 21), under Conclusions.
23
T Wright (1994) Privacy and Electronic Identification in the Information Age (‘With function creep,
systems designed for one purpose are extended over time to other purposes not originally intended’);
SG Davies, ‘Touching Big Brother: How Biometric Technology will Fuse Flesh and Machine’(1994) 7(4)
Information Technology & People 38 (‘The history of identification systems throughout the world pro-
vides evidence of “function creep”–application to additional purposes not announced, or perhaps
even intended, at the commencement of the scheme’).
24
The Google Scholar search (n 20), mentions up to five items yearly in the second half of the 1990s,
rising to over a dozen in 2001 and 2002, further rising from 23 in 2003 through 57 in 2006 to 109
in 2009, and then stabilising at 129 in 2012, 122 in 2015, and 104 in 2018. Note that these
numbers may include a few hits on the physics notion of creep in materials.
25
Collins English Dictionary, http://www.collinsdictionary.com (accessed 15 March 2019).
LAW, INNOVATION AND TECHNOLOGY 7

words, usually without reference to earlier definitions. A representative
sample of definitions is the following:
the addition of new features beyond the scope of the original project;
26
changes in, and especially additions to, the use of a technology;
27
the use of technology for new purposes beyond its originally intended
purposes;
28
the expansion of a program, system or technology into areas for which it was
not originally intended;
29
a process in which agencies use systems for additional purposes that they did
not announce or intend at the beginning of the plan;
30
when a system developed for a particular purpose comes to be used for, or to
provide the underpinnings for other systems that are used for, different
purposes.
31
Whereas these definitions give a rather neutral description of the change in a
system’s function, without referring to how the change comes about, some
authors describe function creep as something that inadvertently happens:
the act of implementing technologies for particular purposes, only to find these
purposes are soon expanded into other unintended areas.
32
Perhaps most important is the issue of ‘function creep.’Once a system has
been developed with a rich set of capabilities, inventive people often can
find other important, beneficial but perhaps also pernicious ways to use it.
(…) Alexander Graham Bell and his early associates never anticipated caller
ID, call forwarding, or automated telemarketing.
33
Perhaps Clarke also saw function creep as something simply happening as
‘additional uses accumulate, and change the purpose of the scheme.’
34
Other authors, however, see function creep as a more or less deliberate
effort by the system’s originators:
26
D Lyon, Surveillance Studies: An Overview (Polity, 2007) 201.
27
Dahl and Saetnan (n 2) 83.
28
A Levin, ‘Big and Little Brother: The Potential Erosion of Workplace Privacy in Canada’(2007) 22(2)
Canadian Journal of Law and Society 197, 212.
29
M Thieme, Privacy Concerns and Biometric Technologies.http://www.bioprivacy.org/privacy_fears.htm.
30
DA Petti, ‘An Argument for the Implementation of a Biometric Authentication System (“BAS”)’(1998) 80
Journal of the Patent and Trademark Office Society 703, 726.
31
MR Curry, DJ Phillips, and PM Regan, ‘Emergency Response Systems and the Creeping Legibility of
People and Places’(2004) 20 The Information Society 357, 362.
32
M Chiesa, R. Genz, F. Heubler, et al., RFID. A Week Long Survey on the Technology and its Potential (2002)
34. https://www.erasme.org/IMG/RFID_research.pdf (accessed 28 December 2020).
33
M Granger Morgan and E Newton, ‘Protecting Public Anonymity’(2004) 21(1) Issues in Science and
Technology 83, 86.
34
(n 21).
8B.-J. KOOPS

‘function’or ‘control creep’(…) describes how a government’s programme of
technological intervention into social life is gradually, incrementally, but delib-
erately, increased over time;
35
databases created for one discrete purpose, despite the initial promises of their
creators, eventually take on new functions and purposes.
36
In a similarly critical vein, several authors, in describing function creep,
stress that the expansion challenges the system’s acceptability:
a process of function creep, [i.e. a process] of the use of information acquired
within the system for other, less universally accepted, purposes;
37
when personal data, collected and used for one purpose and to fulfil one func-
tion, migrate to other ones that intensify surveillance and privacy invasions
beyond what was originally understood, and considered socially, ethically
and legally acceptable.
38
Systems originally intended to perform narrowly specified functions are
expanded to react to new (political) circumstances, thereby sidestepping or
pushing the limits of legal frameworks meant to protect issues of privacy
and data protection.
39
2.4. Conclusion
The definitions of the terms ‘function’and ‘creep’suggest that function creep
of Xhappens when X’s function, i.e. what Xis meant to perform, expands or
shifts slowly, gradually, and/or imperceptibly. It may have a negative associ-
ation when the movement is ‘creepy’, in the sense of being repugnant
(making the flesh creep) or of something sneaky happening, but the creeping
movement might also be experienced more neutrally as just slow, gradual, or
imperceptible.
The definitions of the compound ‘function creep’all share the character-
istic that something’s function is moving beyond its original purpose in a
way that was apparently unforeseen by its developers, users, or the public.
But besides this shared characteristic, the definitions are diverse and empha-
sise different elements. For example, the movement sometimes means that
the original purpose is left behind (indicating a shift), but it also often
35
R. Williams and P. Johnson, Genetic Policing: The Use of DNA in Criminal Investigations (Willan Publish-
ing, 2008) 81–82.
36
T. Simoncelli and B. Steinhardt, ‘California’s Proposition 69: A Dangerous Precedent for Criminal DNA
Databases’(2006) 34 Journal of Law, Medicine & Ethics 199, 283 (emphasis added).
37
(n 31) 359.
38
D. Murakami Wood (ed.), A Report on the Surveillance Society. For the Information Commissioner by the
Surveillance Studies Network (2006).
39
D Broeders, ‘The New Digital Borders of Europe: EU Databases and the Surveillance of Irregular Immi-
grants’(2007) 22(1) International Sociology 22(1) 71, 81.
LAW, INNOVATION AND TECHNOLOGY 9

means that new functions are acquired besides the original purpose (indicat-
ing expansion).
A variety of things can be the object of function creep: a system, technol-
ogy, database, project, or law, typically related to the processing of (personal)
data. The function of the object is often conceived as its purpose(s), although
several authors operationalise the function in terms of its use. The function
of X can thus be defined by its (apparently intended) purpose, but also by its
(factually occurring) usage.
Some stress that the function changed during later development and was
not foreseen by the originators; others stress that the originators (possibly)
knew or even intended the change, but that this was not foreseen by users
or the general public. Function creep thus can be something that befalls X,
but it could also be on someone’s hidden agenda for X.
What is striking in the ‘function creep’definitions in the literature is a
general absence of characteristics associated with ‘creep’. Slowness is not
mentioned in any definition, except being implicitly suggested by Simoncel-
li’s and Steinhardt’s‘eventually’;
40
in contrast, Chiesa and others mention the
expansion happening ‘soon’.
41
Interestingly, Dahl & Saetnan used the title ‘It
all happened so slowly’(carrying the same implicit message as its common
opposite ‘it all happened so quickly’: there was nothing we could do about
it), but do not mention slowness when defining function creep.
42
Only
two definitions mention ‘gradual’or ‘incremental’as a constituent character-
istic.
43
Imperceptibility or stealth are not explicitly mentioned by any
definition, although it can be read between the lines of some (‘only to find
these purposes are soon expanded’,
44
‘despite the initial promises of their
creators’).
45
Creepiness seems at most implicitly included in the definitions
of those authors who critically approach function creep, particularly by
some suggestion of sneakiness in expanding X beyond its originally accepted,
or acceptable, function.
46
Thus, although ‘creep’means a slow, gradual, stealthy, or imperceptible
movement, most definitions of ‘function creep’do not stress any of these
characteristics. Some neutral descriptions carry an implicit suggestion that
the movement was not perceived at the time it occurred but rather after it
happened, but overall, imperceptibility does not seem an essential character-
istic; nor would slowness or gradualness appear crucial for function creep.
Only some critical descriptions contain, albeit rather implicitly, an element
of stealth, with some connotation of sneakiness, arguing that the emergent
40
(n 36).
41
(n 32).
42
(n 2).
43
In Collins Dictionary and Williams and Johnson (n 35).
44
Chiesa et al. (n 32).
45
Simoncelli and Steinhardt (n 36).
46
See, e.g., Simoncelli and Steinhardt (n36) and Broeders (n 39).
10 B.-J. KOOPS

changed functions are less acceptable than the ostensible original one. For
these authors, function creep may well be ‘creepy’. The neutral descriptions,
however, might as well have used a term such as ‘function change’or ‘func-
tion expansion’to indicate the process they are describing. The use of a
specific, more loaded term suggests, nevertheless, that the authors want to
emphasise some mechanism at work that calls for attention. But what that
mechanism is, remains unclear in the literature’s accounts of function
creep. Perhaps a look at concepts closely related to function creep can
help to shed light on these mechanisms.
3. Related concepts
An important part of understanding the concept of function creep is to
analyse how ‘function creep’relates to concepts with which it shares
family resemblances. Placing ‘function creep’within the family of related
concepts helps not only to define its meaning but also to explain its prag-
matic status or linguistic niche in policy debates. It may also help to identify
potential mechanisms underlying the phenomenon of expanding functional-
ity. Various theoretical notions from relevant disciplines, after all, capture
similar notions of new uses of something beyond what was originally fore-
seen or intended.
3.1. Creep concepts
Function creep relates to various other ‘creeps’that describe similar ten-
dencies of things to expand or shift beyond their origins. As mentioned,
Wikipedia –as of August 2019 –automatically refers people searching for
‘function creep’to the page on scope creep:‘changes, continuous or uncon-
trolled growth in a project’s scope, at any point after the project begins.’
47
In
relation to expanding functionality of a product (rather than a project), the
term ‘feature creep’or ‘featuritis’is used.
48
Also, data processing can
expand: data creep is ‘the tendency to continually expand the scope of col-
lection and use of personal information’
49
or ‘the gradual expansion of uses
of information over time’
50
47
https://www.en.wikipedia.org/wiki/Scope_creep (accessed 31 August 2019).
48
‘Feature creep is the excessive ongoing expansion or addition of new features in a product, especially
in computer software, videogames and consumer and business electronics. These extra features go
beyond the basic function of the product and can result in software bloat and over-complication,
rather than simple design’, according to https://www.en.wikipedia.org/wiki/Feature_creep (accessed
31 August 2019).
49
JR Reidenberg, ‘Resolving Conflicting International Data Privacy Rules in Cyberspace’(2000) 52 Stanford
Law Review 1315, 1323.
50
DJ Solove, Understanding Privacy (Harvard University Press, 2008) 131.
LAW, INNOVATION AND TECHNOLOGY 11

At an organisational level, a related term is ‘mission creep’, i.e. the tendency
of organisations to expand their mission or remit.
51
Sometimes, this is even
considered a synonym of function creep: ‘In political science terms, “function
creep”describes the tendencies of bureaucracies to gradually expand their
functions or missions’.
52
This possibly challenges their legitimacy: ‘ICANN’s
nature currently tends to function creep, which emphasizes the need for demo-
cratic legitimacy’.
53
Similarnotionsin relation to public organisations are com-
petence creep
54
and authority creep.
55
Agencies’remit can also expand
through regulation creep,anunderdefined concept denoting the expansion
of regulatory codes,
56
in a way that is uncontrolled
57
or constitutes ‘regulation
by stealth’.
58
Somewhat more broadly in public policy, interest creep denotes
‘the uncritical expansion of underspecified interests like national security and
child protection to capture multiple, distinct sources of government
concern’,
59
which resembles policy stretching (see section 3.4). Similarly, in
the public policy context, but more specifically, the ‘quiet, under-scrutinized,
amorphous expansion of the kinds of information deemed inappropriate for
public consumption’has been termed ‘confidentiality creep’.
60
In Science & Technology Studies, the term ‘surveillance creep’is fre-
quently used.
The surveillance appetite once aroused can be insatiable. A social process of
surveillance creep (and sometimes gallop) can often be seen. Here a tool intro-
duced for a specific purpose comes to be used for other purposes, as those with
the technology realize its potential and ask, ‘Why not?’
61
A related notion is control creep, which
captures the sense in which the apparatus of social control, that is the combi-
nation of technologies and instruments designed to respond to and regulate
deviant behaviour, are becoming increasingly dispersed and interspersed
throughout many different arenas of late-modern social life.
62
51
See. e.g., J Einhorn, ‘The World Bank’s Mission Creep’(2001) 80(5) Foreign Affairs.
52
(n 49), 1323.
53
AM Froomkin, ‘Form and Substance in Cyberspace’(2002) 6 Journal of Small & Emerging Business Law
93, 122.
54
S Weatherill, ‘Competence Creep and Competence Control’(2005) 23 Yearbook of European Law 1.
55
CE Ford and BA Oppenheim, ‘Neotrusteeship or Mistrusteeship? The “Authority Creep”Dilemma in
United Nations Transitional Administration’(2008) 41 Vanderbilt Journal of Transnational Law 55.
56
P Coombes and SC-Y Wong, ‘Why Codes of Governance Work’(2004) (2) The McKinsey Quarterly 48, 52.
57
G Smith, ‘Citizen Oversight of Independent Police Services: Bifurcated Accountability, Regulation Creep,
and Lesson-Learning’(2009) 3(4) Regulation and Governance 421.
58
DT Llewellyn, ‘Trust and Confidence in Financial Services: A Strategic Challenge’(2005) 13(4) Journal of
Financial Regulation and Compliance 333.
59
D Fox, ‘Interest Creep’(2014) 82 George Washington Law Review 273.
60
DS Levine, ‘Confidentiality Creep and Opportunistic Privacy’(2018) Tulane Journal of Technology and
Intellectual Property 20.
61
GT Marx, ‘Seeing Hazily (But Not Darkly) through the Lens: Some Recent Empirical Studies of Surveil-
lance Technologies’(2005) 30 Law and Social Inquiry 339, 385 (emphasis in original, reference omitted).
62
M Innes, ‘Control Creep’(2001) 6(3) Sociological Research Online §2.4.
12 B.-J. KOOPS

All of these describe similar processes as function creep, at varying levels of
abstraction, in relation to policy, organisations, projects, or systems.
However, the term ‘creep’itself seems to be attached to ever more concepts
in widely different contexts –quite fitting for a concept that describes
expanding usage. ‘Christmas creep’, for instance, is a well-established
term to indicate the ‘tendency for Christmas products, decorations and
advertising to be displayed earlier each year’.
63
The inflation of job titles –
‘whereby employers bestow lofty titles on their staffwithout a corresponding
level of authority’–has been termed ‘title creep’.
64
‘Concept creep’is the
tendency to expand the meaning of concepts and applying them to a
broader range of phenomena than before,
65
similar to the inflation of
terms in ‘terminological creep’.
66
And recently, Greenleaf coined ‘GDPR
creep’to indicate that ‘[c]ompanies outside Europe are adopting “GDPR
compliance”across their whole business operations, even though there is
no legal obligation to do so’.
67
I could extend this list further, but it seems
more appropriate instead to introduce the term ‘creep creep’here: the
inflationary tendency, outside the context of technologies, projects,
systems, or organisations, to designate the expansion of X beyond its original
situation with the term ‘Xcreep’.
3.2. STS concepts
In Science & Technology Studies (STS), function creep relates to Langdon
Winner’s notion of reverse adaptation: the adjustment of human ends to
match the character of the available means.
68
This describes the push of tech-
nology systems for function expansion, where technology systems indicate
‘large sociotechnical aggregates’that include all humans associated with
the system.
69
‘Technical systems become severed from the ends originally
set for them and, in effect, reprogram themselves and their environments
to suit the special conditions of their own operation.’
70
Winner’s theory of
technological politics advanced the hypothesis that ‘as large-scale systems
come to dominate various areas of modern social life, reverse adaptation
63
https://www.macmillandictionary.com/buzzword/entries/Christmas-creep.html (accessed 31 August
2019). See also https://www.en.wikipedia.org/wiki/Christmas_creep (accessed 31 August 2019).
64
MLaff,‘Beware of Title Creep’(2008) T+D(July) 20.
65
N Haslam, ‘Concept Creep: Psychology’s Expanding Concepts of Harm and Pathology’(2016) 27 Psycho-
logical Inquiry 1.
66
D Beetham, ‘The Right to Development and Its Corresponding Obligations’in BA Andreassen and SP
Marks (eds), Development as Human Right: Legal, Political, and Economic Dimensions (Harvard School of
Public Health, 2006) 79.
67
G Greenleaf, Global Convergence of Data Privacy Standards and Laws. Speaking Notes for the European
Commission Events on the Launch of the General Data Protection Regulation (GDPR) in Brussels & New
Delhi, 25 May 2018, 3. https://www.papers.ssrn.com/sol3/papers.cfm?abstract_id=3184548.
68
L Winner, Autonomous Technology (MIT Press, 1977) 227–9.
69
Ibid., at 242.
70
Ibid., at 227.
LAW, INNOVATION AND TECHNOLOGY 13

will become an increasingly important way of determining what is done and
how’.
71
The prevalent use of the term function creep in the past three decades
to criticise all kinds of public-policy systems and initiatives seems to corro-
borate Winner’s hypothesis. This could also be because technology may
induce the classical moral maxim ‘ought implies can’(a duty only exists if
an action is doable) to flip around: ‘can sometimes implies ought’, because
‘by creating new practical opportunities, technology tends to create new obli-
gations and rights’.
72
Before Winner, Martin Heidegger had already articulated the force of
technology in influencing how we perceive the world in terms of means
and ends. Technology is not merely an instrument under human control,
but a way of revealing (Entbergen) the world, i.e. making us see things in
certain ways. In particular,
[t]he revealing that rules throughout modern technology has the character of a
setting-upon, in the sense of a challenging-forth. That challenging happens in
that the energy concealed in nature is unlocked, what is unlocked is trans-
formed, what is transformed is stored up, what is stored up is, in turn, distrib-
uted, and what is distributed is switched about ever anew.
73
In a very abstract sense, this may help explain mechanisms of function creep,
as modern technology makes us see everything as further resources to be
used and manipulated. This enframing (Ge-stell) of modern technology
entails ordering everything in such a way that it can be used most effectively
and efficiently, hiding alternative ways of seeing and using the world.
74
Thus,
the tools and systems that we build are not so much under our control but
expressions of this all-encompassing enframing.
75
Similarly but less determi-
nistically, Verbeek’s articulation of technological intentionality shows how
‘[t]echnologies help to shape actions because their scripts evoke given beha-
viors and because they contribute to perceptions and interpretations of
reality that form the basis for decisions to act’.
76
In that sense, technologies
invite humans to use them in certain ways to do things, in ways unforeseen
without these technologies’directing human action and experience.
77
Somewhat more concretely closer to function creep is Jacques Ellul’snotion
of self-augmentation (auto-accroissement), which he saw as a key characteristic
of ‘technique’(a broader concept than technology, referring to all methods or
means to achieve an end in any field of human activity). Ellul conceives self-
71
Ibid., at 251.
72
T Swierstra, ‘Nanotechnology and Technomoral Change’(2013) 15(1) Etica and Politica 200, 210.
73
M Heidegger, The Question Concerning Technology, and Other Essays (Harper & Row, 1977) 16.
74
Ibid., at 19–20.
75
Ibid., at 20–21.
76
P-P Verbeek, Moralizing Technology: Understanding and Designing the Morality of Things (University of
Chicago Press, 2011) 56.
77
Ibid., at 57.
14 B.-J. KOOPS

augmentation as ‘the fact that everything happens as if the technical system
grows through an internal force, intrinsically and without decisive human inter-
vention’.
78
Each technical invention provokes other inventions in other
domains, in a never-ending process (although Ellul acknowledges that, while
the whole of techniques never stops developing, individual techniques may
face barriers impeding progress).
79
The main mechanism is that when a new
technical form appears, this permits and conditions various other forms; as
soon as someone discovers a technical process, people note how it can be
applied in many other domains than the one for which it was invented.
80
The vision of technology as an autonomous force, almost or completely out
of control by humans, has long been criticised in STS literature. Many studies
have shown the numerous ways in which humans actually considerably
influence technology development (intentionally or non-intentionally).
Feature creep and scope creep are examples of mechanisms whereby designers
and project managers or stakeholders push for the addition of new function-
alities along the way. Also, users frequently find ways to use applications in
ways completely unforeseen or unintended by their developers or marketers.
81
Nevertheless, the notions of enframing, self-augmentation, and reverse adap-
tation still tell us something about a possible mechanism at work in function
creep: when the need arises to solve a problem, people logically look at existing
solutions and how they could be adapted in light of new circumstances. In that
sense, existing technologies or systems ‘beg’to be used in new contexts,
‘because it would be folly not to use the means’.
82
3.3. Economics concepts
The phenomenon of gradual expansion of functionality, purpose, or use has
obvious benefits. In fact, function creep can also be seen as a form of inno-
vation, that is, ‘the application of new ideas to the products, processes, or
other aspects of the activities of a firm that lead to increased “value”’.
83
Using a system for a purpose not originally foreseen is applying a new
idea to it, namely to use an existing system in a new way. Innovation is
not a one-offactivity, but an iterative process, which is driven by ‘the percep-
tion of a new market and/or new service opportunity for a technology-based
invention’combined with efforts aimed at market introduction.
84
In
78
J Ellul, Le Système technicien (Calmann-Lévy, 1977) 229 (my translation).
79
J Ellul, La technique ou l’enjeu du siècle (Paris: Economica, 1990) 83.
80
Ellul (n 79) at 81.
81
Oudshoorn and Pinch (n 5).
82
Ellul (n 78).
83
C Greenhalgh and M Rogers, Innovation, Intellectual Property, and Economic Growth (Princeton Univer-
sity Press, 2010) 4.
84
R Garcia and R Calantone, ‘A Critical Look at Technological Innovation Typology and Innovativeness
Terminology: A Literature Review’(2002) 19 The Journal of Product Innovation Management 110, 112.
LAW, INNOVATION AND TECHNOLOGY 15

organisational terms, innovation refers to ‘the multi-stage process whereby
organizations transform ideas into new/improved products, service or pro-
cesses, in order to advance, compete and differentiate themselves successfully
in their marketplace’.
85
Similar to public policy (section 3.4), innovation can
be incremental or evolutionary, building up over time, but also discontinu-
ous or ‘revolutionary’, and it may be disruptive when steady improvements
by (initially fringe) firms build up to outperform and displace existing
market leaders.
86
The definitions of innovation suggest it is an obvious complement to, or
mirror of, the reverse adaptation of technology, which explains the expan-
sion of a technology’s functionality not so much as a natural tendency of
technology to remain relevant in new circumstances, but as a natural ten-
dency of businesses to keep competing in changing markets. With
different emphasis as to the primary driver, both concepts capture some of
the logic underlying new uses of a system. Calling such new use ‘innovative’
is to emphasise its positive value, in economic or social terms. Calling such
new use ‘function creep’is to emphasise negative aspects that often come
along with positive benefits, including negative externalities, that is, costs
borne by others that are not internalised in the cost/benefit analysis of the
one deciding on a certain activity.
3.4. Policy concepts
In policy studies, function creep relates to the notion of incrementalism: the
development of public policy in small increments. The development of
public-policy alternatives and the enactment of law is often ‘quite small,
gradual, and incremental’.
87
Lindblom observed that for complex problems,
administrators rely principally on the ‘method of successive limited compari-
sons’, developing policy step by small step.
88
This can lead to policy stretch-
ing,if‘operating over periods of decades or more, elements of a [policy] mix
are simply extended to cover areas they were not intended to at the outset’.
89
Policy stretching is particularly problematic if the elements ending up in the
mix incorporate contradictory goals or instruments, frustrating initial policy
goals; further tinkering to address this may even compound the problems.
90
85
A Baregheh, J Rowley, and S Sambrook, ‘Towards a Multidisciplinary Definition of Innovation’(2009) 47
(8) Management Decision 1323, 1334.
86
CM Christensen, The Innovator’s Dilemma: When New Technologies Cause Great Firms to Fail (Harvard
Business School Press, 1997).
87
JW Kingdon, Agendas, Alternatives, and Public Policies (Harper Collins, 1995) 83.
88
CE Lindblom, ‘The Science of “Muddling Through”’ (1959) 19(2) Public Administration Review 79, 88.
89
M Howlett and I Mukherjee, ‘Policy Design and Non-Design: Towards a Spectrum of Policy Formulation
Types’(2014) 2(2) Politics and Governance 57, referring to PH Feindt and AG Flynn, ‘Policy Stretching
and Institutional Layering: British Food Policy between Security, Safety, Quality, Health and Climate
Change’(2009) 4 British Politics 386.
90
Howlett and Mukherjee (n 89) 63–64.
16 B.-J. KOOPS

Sometimes, however, policy can also change quickly, disrupting the status
quo until a new equilibrium is reached; this process of punctuated equili-
brium might resemble instances of function creep where several small
steps build into what is suddenly perceived as a large step that significantly
altered the status quo.
91
A notion (initially from economics) close to incrementalism is the ratchet
effect,‘the well-known tendency of planners to use current performance as a
criterion in determining future goals’.
92
This effect is also associated with an
incremental, possibly slow, development, as targets in the following year are
based on performance in the current or previous year; this has a perverse
incentive for performers not to outperform their current target since this
will have knock-on effects in future years. In the context of public institutions,
a ratchet effect has also been pointed out in the form of a gradual increase in
governmental authority over resource allocation, where authority is expanded
in periods of crisis but left in place after the crisis has abated (possibly limited
somewhat, but with a still greater scope than before the crisis).
93
Although he
based his analysis on pre-1987 cases, Higgs observes in his 2012 preface that
the ‘logic of the ratchet effect remains applicable today as at any time in the
past century.’
94
This so-called Higgsian ratchet is one instantiation of
mission creep, and can be explained by the observation that,
[o]nce undertaken, governmental programs are hard to terminate. Interests
become vested, bureaucracies entrenched, constituencies solidified. More fun-
damentally, each time the government expands its effective authority over
economic decision-making, it sets in motion a variety of economic, insti-
tutional, and ideological adjustments whose common denominator is a dimin-
ished resistance to Bigger Government.
95
Function creep is more than just incrementalism, however. It rather
resembles Kingdon’s notion of policy spillover: policy success in one area
contributes to success in adjacent areas.
96
This occurs
because politicians sense the payoffin repeating a successful formula in a
similar area, because the winning coalition can be transferred, and because
advocates can argue from successful precedent. These spillovers are extremely
powerful agenda setters, seemingly bowling over even formidable opposition
that stands in the way.
97
91
FR Baumgartner and BD Jones, Agendas and Instability in American Politics (University of Chicago Press,
1993).
92
ML Weitzman, ‘The “Ratchet”Principle and Performance Incentives’(1980) The Bell Journal of Economics
302, at 302.
93
R Higgs, Crisis and Leviathan: Critical Episodes in the Growth of American Government (The Independent
Institute, 1987, 2012).
94
(n 93), at xvi.
95
(n 93) at 261.
96
Kingdon (n 87), at 190–194.
97
Ibid., at 203.
LAW, INNOVATION AND TECHNOLOGY 17

Spillovers occur not only within policy domains or subsystems (when a
formula is repeated through precedent) but also across policy domains,
when ‘a precedent spills over from one arena into an adjacent one’.
98
While the force of policy spillovers helps explain the potential of function
creep to occur, two factors are relevant that could counter function creep
from happening. First, there is a short time-frame for using the power of pre-
cedent: the ‘[policy] window in the first area opens windows in adjacent
areas, but they close rapidly as well’; problems will inevitably show when
the first policy is being implemented, limiting the attractiveness of spil-
lovers.
99
Second, using the argument from precedent requires careful cat-
egory construction, since both policy areas need to be presented as similar
for the solution in the first to work in the second.
100
This implies that,
from a public policy perspective, function expansion is more likely to
occur in connected, similar contexts (for instance, from national security
to law enforcement) than between completely different contexts (such as
from health to law enforcement).
3.5. Sociological concepts
The ratchet effect has a broader application beyond economics and policy. In
a broad sociological sense (drawing on insights from evolutionary anthropol-
ogy and cognitive psychology), the ratchet effect has also been articulated in
how human culture develops.
One generation does things in a certain way, and the next generation then does
them in that same way –except that perhaps they add some modification or
improvement. The generation after that then learns the modified or improved
version, which then persists across generations until further changes are
made.
101
This ratchet effect characterises human cultural transmission, whereby
‘modifications and improvements stay in the population fairly readily
(with relatively little loss or backward slippage) until further changes
ratchet things up again’.
102
Culture develops cumulatively through inventiveness (causing modifi-
cations) and transmission (causing the modification to spread and
sustain). Tennie et al. point out three mechanisms that are fundamental to
human transmission: teaching, social imitation, and normativity. Particu-
larly the latter is relevant from the perspective of function creep. In teaching
98
Ibid., at 190.
99
Ibid., at 192.
100
Ibid., at 193.
101
C Tennie, J Call, and M Tomasello, ‘Ratcheting Up the Ratchet: On the Evolution of Cumulative Culture’
(2009) 364 Phil Trans R. Soc B 2405.
102
Ibid.
18 B.-J. KOOPS

and social imitation, children not only learn how something is often done,
but also apparently understand that this is ‘rather the way it should be
done. (…) [Y]oung children learn very quickly that a particular artefact is
“for”a particular function, and its other uses may be considered
wrong.’
103
Cultural transmission, therefore, has an innate tendency of con-
servatism –conformity with how things are traditionally done –which
can be evolutionarily explained by its guaranteeing ‘an unusual degree of
faithful transmission across generations (…) in a way that supports the
further ratcheting up in complexity of cultural artefacts and practices
across historical time.’
104
Thus, the cultural-evolutionary ratchet effect not
only reflects the notion of incrementalism (change must be slow and
gradual in order to be acceptable), but it also sheds some light on the ‘creepi-
ness’of function expansion. Novel usage of an artefact triggers resistance
because this is not what the artefact is ‘for’, and doing something different
than what people are used to feels ‘wrong’in a strongly rooted evolutionary
psychological sense.
3.6. Legal concepts
The notions of mission creep, competence creep, and authority creep res-
onate in the legal concept of abuse of power or détournement de pouvoir.
This designates exercise of authority in the public interest but not for the
precise purpose for which it was originally granted.
105
Although normally
well-intentioned, it is problematic becauseifapowergrantedforpurpose
Aisusedforadifferent purpose B, this will usually create a legitimacy
deficit.Thereisnoexplicitlegalbasisforhowthepowerisused,and
thelegislatorhasnotconsideredthespecific checks and balances that
might be needed for using the power for purpose B (or C, etc.) rather
than A.
The deviation from original purposes is also a well-known phenomenon
in the processing of personal data, and data protection law has developed a
compatibility test to keep secondary use of personal data in check. Per-
sonal data can be collected only for specified, explicit, and legitimate pur-
poses; they can be further processed, but only if this is not incompatible
with those purposes (art. 5 General Data Protection Regulation). Authors
frequently use the term ‘function creep’in this context, to refer to unac-
ceptably deviating forms of processing. For example, Kindt describes func-
tion creep as ‘the risk that the data are used for secondary purposes which
are not compatible with the purposes for which the data were initially
103
Ibid., at 2412 (emphasis in original).
104
Ibid.
105
B Horvath, ‘Rights of Man. Due Process of Law and Excès de Pouvoir’(1955) 4 The American Journal of
Comparative Law 539, 567–568.
LAW, INNOVATION AND TECHNOLOGY 19

collected.’
106
Wisman originally used ‘function creep’for the ‘use of data
for a different goal than it was collected for’,
107
but (following a suggestion
by Arno Lodder) eventually termed this ‘purpose creep’. Similarly,
Jentzsch describes purpose creep as ‘the tendency to use information for
more and more purposes –also those that are unrelated to purposes for
which the information was originally collected.’
108
Secondary use also resonates with a notion from a different legal context:
transformative use, which refers to copyrighted material being quoted ‘in a
different manner or for a different purpose from the original,’which is a rel-
evant factor for claiming the fair-use exception in copyright law.
109
3.7. Discourse concepts
The precedent element in the concept of policy spillover ties in with a notion
in argumentation theory: the slippery slope argument. This type of argu-
ment holds that if you take the seemingly desirable step A, through some
process you will end up taking the undesirable step B. Conceptual and
empirical studies suggest that slippery slope arguments are not always falla-
cious: they can be used –in critical discussions, particularly in institutiona-
lised contexts such as law –to shift a burden of proof
110
or to bring the
discussion to a higher level by looking at the cumulative effect of many mar-
ginal decisions.
111
A slippery slope argument can be valid, if it not only
explains why B is undesirable, but also plausibly argues why A will lead to
B. Possible mechanisms for A leading to B are that A lowers the costs for
B, that A changes attitudes, power structures, or political momentum to
favour B, and that people’s tolerance for small changes will disarm resistance
to (incremental steps towards) B.
112
Coincidentally, the metaphor of a slippery slope resounds in one of the
more technical meanings of ‘creep’, namely a ‘slow, imperceptible movement
en masse of soil, talus, etc., usu. downhill under the influence of gravity but
freq. with other processes (such as successive freezing and thawing)
106
E Kindt, ‘Biometric Applications and the Data Protection Legislation. The Legal Review and the Pro-
portionality Test’(2007) 31(3) Datenschutz und Datensicherheit 166, 167; similarly, MS de Vries, ‘Hoe
Waarschijnlijk is Function Creep? Een beleidswetenschappelijke analyse’(2011) 37(8) Justitiële verken-
ningen 22, 23.
107
THA Wisman, ‘Purpose and Function Creep by Design: Transforming the Face of Surveillance Through
the Internet of Things’(2013) 4(2) European Journal of Law and Technology s. 1.
108
N Jentzsch, Financial Privacy: An International Comparison of Credit Reporting Systems (Springer, 2007)
139.
109
PN Leval, ‘Towards a Fair Use Standard’(1990) 103 Harvard Law Review 1105.
110
W Van der Burg, ‘Slippery Slope Arguments’in R. Chadwick (ed.), International Encyclopedia of Applied
Ethics (Academic Press, rev 2nd edn 2012) 129, 139; DN Walton, Slippery Slope Arguments (Oxford Uni-
versity Press, 1992) 13.
111
MJ Rizzo and DG Whitman, ‘The Camel’s Nose in the Tent: Rules, Theories, and Slippery Slopes’(2003)
51 UCLA Law Review 539.
112
E Volokh, ‘The Mechanisms of the Slippery Slope’(2003) 116 Harvard Law Review 1026.
20 B.-J. KOOPS

contributing to the effect’.
113
To be sure, the metaphor suggests that it is
humans who will slide down the slope by making a first step downhill, not
the earth itself, but there is a serendipitous connection between the image
of an earth mass creeping downhill and humans sliding downhill in a
process of function creep.
In the literature, authors frequently point to a risk of function creep in a
manner that suggests a slippery slope argument, particularly when criticising
a proposal for a new law, policy, or database. For instance, Simoncelli and
Steinhardt argue that ‘[c]ontinuing down a path of unaccountable “function
creep”may bring us to a day when the entire U.S. population finds itself in a
government [DNA] database.’
114
Such an argument usually makes clear why
the (future) use of the proposed system for different purposes is undesirable,
but authors do not always (explicitly) make a case for why the functionality
of the proposed system is likely to expand in the future. Sometimes, there is a
suggestion of a hidden agenda (namely that the proponents of the object
already have different future uses in mind), but authors also frequently
seem inclined to follow Ellul’s assumption of self-augmentation and
Winner’s hypothesis of reverse adaptation, considering the future expansion
too obviously likely to have to explain the pathway of future expansion. From
an argumentation-theoretical perspective, it is questionable whether this
results in a successful slippery slope argument and a fruitful debate. Propo-
nents could easily negate that there is a risk of function creep, by pointing out
that future different uses are not envisioned and that expansion of function-
ality can be discussed in future when some occasion for expansion would
arise.
A related type of argument is the argument from added authority,
115
i.e.
an argument ‘holding that someone should not be given a certain authority or
responsibility because he will probably abuse it’.
116
This argument ‘proceeds
from the assumption that even the unlikely becomes more likely once jurisdic-
tion is granted than it would have been without that jurisdiction.’
117
Thus, it
‘cautions against granting jurisdiction for fear that the jurisdiction, once
granted, will be available to decide some possible future case in some way
admittedly feared by the decisionmaker as well as by the maker of the argu-
ment’.
118
Although similar, according to Van der Burg, it is not a slippery
slope argument, but rather an argument pointing out the risk of future
abuse of power –possibly in the strict sense of détournement de pouvoir,if
the power were used for purposes for which it was not established, but also
113
OED (n 3).
114
Simoncelli and Steinhardt (n 36) at 290.
115
F Schauer, ‘Slippery Slopes’(1995) 99(2) Harvard Law Review 361, 367–8.
116
W Van der Burg, ‘Slippery Slope Arguments’in R Chadwick (ed.), Encyclopedia of Applied Ethics (Aca-
demic Press, 1998) 129.
117
Schauer (n 115) at 368.
118
Ibid., at 367.
LAW, INNOVATION AND TECHNOLOGY 21

possibly within the scope of the authority granted but applied in a context
where it might have undesirable consequences.
119
If a court were, for
example, to assume the authority to ban racist advocacy, based on the argu-
ment that racist ideas are wrong, courts would then acquire an accepted
added authority to decide which ideas are wrong, and might also apply it to,
say, anti-war or socialist ideas.
120
Function creep arguments sometimes take
this form, when pointing out the risks of investing the government with a
new power. Similarly, arguments of interest creep and surveillance creep
can also function to point out risks of investing an institution with added auth-
ority. Moreover, the argument has strong similarities with the way authors fre-
quently use the term ‘function creep’to warn against creating a new database
or system: once the database or system is there (even if innocuous in its pro-
posed form), there is an increased likelihood that the database or system will
be used for undesirable purposes, even if such use currently seems unlikely.
3.8. Conclusion
The wide, expanding usage of ‘function creep’, other ‘creeps’, and related
concepts suggests that the phenomenon of gradual expansion beyond what
was originally foreseen is a widespread, perhaps natural, phenomenon in
social systems. The analysis of related concepts helps to understand possible
mechanisms underlying this phenomenon.
Function creep can be understood by situating it in the nexus where many
related theoretical notions meet, most notably reverse adaptation and self-
augmentation of technology systems, incrementalism and disruption in
policy and innovation, policy spillovers, ratchet effects, transformative use,
and slippery slope argumentation. Technological systems have a tendency
to reprogramme themselves and their environments to adapt to the evolving
conditions of their own operation. Businesses have an incentive to innovate
by finding new uses of existing technology, and policy-makers have a ten-
dency to copy proven solutions and apply them in other contexts. Gradual
expansion or incrementalism is a widespread phenomenon in social
systems, where ratchet effects ensure that a novelty is taken up, integrated
in the status quo, and therewith normalised, with little or no likelihood of
slipping back to the previous situation. These incremental changes may
accumulate to effect transformative change, but the transformative effect
may be unnoticed or unacknowledged, leading opponents to use slippery-
slope arguments to contest the change. Thus, this complex of notions
bearing significant family resemblances sheds light, from different angles
highlighting different mechanisms, on why and how function creep occurs.
119
(n 116) at 131.
120
Volokh (n 112) at 1065–1066.
22 B.-J. KOOPS

4. Defining function creep
In light of the above analysis, how can we define function creep? Several
components of the definition are relatively straightforward. Function creep
mainly refers to something happening (as in the verb ‘creep’), typically an
expansion but sometimes rather a shift. What changes is the function of
something, that is, the regular or proper activity of something in relation
to its purpose and/or use. This change is often intentional (although
usually not intentionally ‘creepy’), but can also happen inadvertently, so
that the definition should not include intentionality.
The something undergoing function creep is typically a system, technol-
ogy, database, project, or law, almost always related to the processing of
(personal) data. Since databases are (part of) a system, and function-creep-
ing technologies, projects, laws typically involve some data processing
system, ‘data-processing system’suffices for definition purposes. The litera-
ture mentions ‘function creep’particularly when public-sector information
systems are used for new functions, frequently in other contexts, most
notably in surveillance, anti-terrorism, e-government, and e-health policies,
often in connection to identification schemes or databases. A sharp
definition might summarise this as public-sector data-processing systems,
but that may be too narrow since also private-sector systems or public-
private partnerships might be involved, depending on the national-insti-
tutional context.
Less straightforward is defining the ‘creeping’(or perhaps ‘creepy’)
element of function creep. Existing definitions of ‘function creep’vary con-
siderably as to the key characteristic of a ‘creeping’movement: is it slow,
gradual, stealthy, or imperceptible? Slowness and gradual movement res-
onate with the incrementalism of public policy and innovation, and more
generally with the small steps visible in ratchet effects of many social
systems’learning and development cycles. But ‘function creep’is also used
for relatively quick or large changes (as in surveillance ‘gallop’). Gradualness
seems characteristic of many cases, suggesting several small steps, but func-
tion creep can also refer to single, sometimes one-off, changes. Stealth, or
even sneakiness, is sometimes invoked in critical narratives, but not in the
majority of the literature. Neither does imperceptibility seem an essential
characteristic in most authors’accounts of function creep.
Yet there must be some ‘creep’element that is relevant to invoke. The
negative connotation that ‘creep’clearly has, in function and other creep
concepts, has to be significant. Otherwise, authors could simply speak of
‘function change’or ‘innovation’(because using something for a new
purpose is applying a new idea to it, which is what innovation basically
entails). Neutral definitions or descriptions, which are surprisingly prevalent
(see section 2.3.2), are therefore missing a point. But which point, exactly?
LAW, INNOVATION AND TECHNOLOGY 23

I suggest that the ‘creep’element in ‘function creep’can best be captured
by focusing on imperceptibility –which might be caused by slowness, gra-
dualness, or stealth –in a specific sense. Namely, the imperceptibility of
the tipping point between linear and non-linear change, which is difficult
to pinpoint. This imperceptibility is an issue, because the difference
between linear and non-linear change is often –at least in the context of
function creep –connected to a difference between acceptable/non-contro-
versial and unacceptable/contested change. This suggestion is based on the
insight that literature on innovation and public policy (among other
domains) generally distinguish between linear, gradual change and non-
linear, disruptive change (cf. sections 3.3–3.4). This is mirrored in the
notion of ‘transformative use’(section 3.6), which denotes some qualitative
difference between old and new usage. Often, it is not evident whether some
change is gradual or disruptive. Sometimes, we only note in hindsight that
what seemed incremental change was actually disruptive (or vice versa).
More often, I think, gradual change at some point simply accumulates
into something qualitatively different, and then proves disruptive. As a mole-
hill at some point becomes a mountain if enough sand is added, so does a
forensic DNA database of offenders at some point turn into a nation-wide
database if enough people (suspects, family, dragnet volunteers) are added.
In this process, there must be a tipping point where quantitative change
becomes qualitative change; but it is very hard to pinpoint where or when
exactly this occurs.
This imperceptibility of the tipping point is what is subtly emphasised
when someone labels something ‘function creep’. If the label is applied ex
ante, this indicates that it is foreseeable that a proposed system will
expand in the future; since the expansion may escape notice because it con-
sists of small, gradual steps and the point at which such expansion becomes
unacceptable –the tipping point –is hard to define or predict, we had better
discuss the possible expansion now and take measures to prevent the system
getting out of hand. If the label is applied ex post, this indicates that expan-
sion occurred over time but was not perceived at any point to create a quali-
tative difference that makes it, in hindsight, contested; so that now we have a
problem (because ratchets are as hard to push back as genies in bottles). The
label can also be applied ex nunc, as an argument against a particular expan-
sion proposed at the moment. In this case, it is clear to opponents that the
tipping point is now and that the proposed change disrupts the status quo;
but the proponents do not see this, arguing that the change is linear
(fitting the system’s purpose) and therefore acceptable. In such cases, the
tipping point is imperceptible in that its existence is being unacknowledged
by the proposal’s proponents.
Now, imperceptibility as such is not necessarily negative. However,
imperceptibility of a tipping point between linear/acceptable and non-
24 B.-J. KOOPS

linear/unacceptable change is problematic. This, I argue, is what the ‘creep’
element in ‘function creep’(and in many other creep concepts) denotes. The
new function is, in some sense, unacceptable (to the one speaking of function
creep), but the acceptability of the change can (or could) not be discussed
properly because the change is (or will be or was) not generally perceived
to be controversial at the material time. Why a function expansion or shift
is unacceptable (for opponents), can vary widely, from making the system
ineffective or unmanageable (e.g. because of featuritis, over-complexity, or
contradictory rules) to negative externalities, such as loss of human
control (e.g. through self-augmentation or reverse adaptation), a legitimacy
deficit (e.g. through mission creep, abuse of power, or incompatible second-
ary use), or lack of appropriate checks and balances (e.g. because a system
expands to other domains with different norms). But whatever the reason,
that a function change is, in some sense, problematic because of unperceived
or unacknowledged transformative use, is what the term ‘function creep’
aims to highlight.
This brings me to propose a definition of function creep that reflects the
above analysis. Function creep comes down to an expansion or shift in the
proper activity of a data-processing system in relation to its purpose or
use, which is considered contestable because people insufficiently recognise
that it involves a qualitative change that raises public-policy concerns.
Although that might serve as a precise definition, it is rather long, and
since definitions are preferably succinct and limited to essential elements, I
propose the following, shorter, definition: function creep denotes an imper-
ceptibly transformative and therewith contestable change in a data-pro-
cessing system’s proper activity.
Key elements in this definition are ‘imperceptibly’and ‘therewith’. These
specify that the change is contestable because it is imperceptibly transforma-
tive, that is, because it is transformative in a way that is imperceptible. This is
what distinguishes function-creep changes from (merely) innovative
changes. Incremental innovation can also, at some point, transform into dis-
ruptive innovation, and the tipping point may be equally imperceptible;
while such transformation may easily raise public-policy concerns (as dis-
ruptive innovation often does) and thus be contestable, this is not because
the transformation was imperceptible or unperceived, but because of the
transformation itself and its consequences. Function creep, in contrast,
denotes some qualitative change that causes concern not only, and I
suggest not primarily, because of the transformation and its consequences
as such, but rather because the transformation is or was not perceived at
the time to be qualitative and in need of discussion. In other words, what
the ‘function creep’label crucially highlights is that there is or has been
insufficient room for debate about the acceptability of the change. What is
contested is not (only) that a change is transformative, but (also) that it is
LAW, INNOVATION AND TECHNOLOGY 25

imperceptibly transformative. In debates, the ‘function creep’label therefore
illuminates that there is some qualitative change at issue that was hitherto
unrecognised, and simultaneously calls for the much-needed debate about
this qualitative change that should have taken place earlier, if only the
actor(s) responsible for the system had realised, or acknowledged, at the
proper time that the change was transformative.
Therefore, I suggest that ‘function creep’is pejorative not, or not primar-
ily, because some change in a system is unacceptable per se, but because the
change is not properly debated, and this lack of a proper debate as much as
the transformation itself has undesirable consequences (such as a legitimacy
deficit, insufficient checks and balances, or a poorly functioning system).
This is not to suggest that having a proper debate will usually suffice;
often, of course, measures will be needed to resolve the public-policy
issues triggered by the transformative change, and perhaps the change
should be stopped or turned back altogether. But at least a proper debate
should help in identifying problems and possible solutions, and while the
outcome will not satisfy everyone, the debate should at least remove the cree-
piness from the system’s expanding functions.
I am also not suggesting that people use ‘function creep’only to highlight
the lack of a proper debate; often, they also want to highlight the negative
consequences of the function expansion as such. Indeed, most authors use
the term ‘function creep’in passing and spend most of their paper discussing
what is wrong with a contested system itself. However, the framing of a
problem matters, because it influences the direction in which solutions are
sought.
121
It is therefore important, in debates about a system’s function
expansion from A to B, to distinguish between criticising:
(1) B as such (implying the system should never be made to do B);
(2) the move from A to B (implying this move should only be done, if at all,
under certain, as yet unfulfilled, conditions); and
(3) how the move from A to B is taking place (implying the move should be
approached differently).
Often, (2) and (3) will go together, because substance (2) is often inter-
twined with process (3). However, criticism (2) is, ultimately, really a ‘func-
tion change’or a ‘disruption’argument; only (3) can properly be called a
‘function creep’argument. People who criticise any move from (original,
acceptable) A to (new, unacceptable) B as a form of ‘function creep’are over-
stretching the meaning of the term –unless they are also explicitly arguing
121
DA Schön, ‘Generative Metaphor: A Perspective on Problem-Setting in Social Policy’in A Ortony (ed.),
Metaphor and Thought (Cambridge University Press, 2nd edn 1993) 137.
26 B.-J. KOOPS

that the move is wrong because it is imperceptible (through slowness, gra-
dualness, or stealth) and therefore receives insufficient attention.
5. Conclusion
In this paper, I have analysed the concept of function creep in order to
develop a definition of this widely used but under-researched term. Based
on semiotic analysis of the term ‘function creep’and its constituent parts,
and of closely related concepts from different fields, I have argued that
what is distinctive of function creep (as opposed to simply function expan-
sion or innovation) is that it denotes some qualitative change in a system’s
function that causes concern not only because of the change itself but also
because the change is insufficiently acknowledged as being transformative
and in need of discussion. Therefore, I proposed the following definition:
function creep denotes an imperceptibly transformative and therewith con-
testable change in a data-processing system’s proper activity.
122
Gradual expansion beyond what was originally foreseen is a widespread,
perhaps natural, phenomenon in social systems. In that light, function creep
can best be understood by situating it in the nexus of many related theoreti-
cal notions, most notably reverse adaptation and self-augmentation of tech-
nology systems, incrementalism and disruption in policy and innovation,
policy spillovers, ratchet effects, transformative use, and slippery slope argu-
mentation. These theoretical notions help us to understand that transforma-
tive function expansion is hard or impossible to prevent; it is ingrained in
many social systems’development cycles. Dealing with function expansion
in data-processing systems that qualitatively changes the status quo and
raises normative issues (e.g. legitimacy deficits or inadequate checks and bal-
ances), is a daunting task for regulators. Likely, they need to resort to
dynamic, self-learning approaches such as responsive regulation
123
and
stimulate responsible innovation.
124
That, however, is not the same as dealing with function creep. As I have
argued, function creep is primarily a notion that captures how transformative
change occurs, namely under the radar. It is a concept used in debates to
highlight the imperceptible (because slow, gradual, or stealthy) way in
which a system’s function changes. Therewith, it functions first and foremost
122
Although not the primary purpose of this paper, defining function creep in this way also enables
giving a generic definition of creep concepts in general. Given the close family resemblances
between prevalent ‘X creep’terms, we can see the term ‘creep concept’as the parent (or umbrella)
concept of this family. A ‘creep concept’is a concept in the form of ‘X creep’that denotes an imper-
ceptibly transformative and therewith contestable change in X.
123
I Ayres and J Braithwaite, Responsive Regulation: Transcending the Deregulation Debate (Oxford Univer-
sity Press, 1992).
124
R Owen, J Bessant, and M Heintz, Responsible Innovation: Managing the Responsible Emergence of
Science and innovation in Society (Wiley, 2013).
LAW, INNOVATION AND TECHNOLOGY 27

as an argument in debates to make visible that something is going wrong, and
therewith to shift the burden of proof back to those responsible for a system
to do something about this. In that sense, function creep is easier to deal with
than transformative function change as such. It can be prevented by regularly
scrutinising whether a data-processing system’s functioning is, over time or
otherwise imperceptibly, qualitatively changing, and as soon as that seems
the case, to be open about this and discuss it. And if the transformative
change has already happened and opponents use a function creep argument,
those responsible for the system can take up the challenge and argue why
they think the change is in line with what the system is supposed to do,
and where needed, take appropriate measures to make it more acceptable.
Acknowledgements
I thank Maša Galič, Martin Husovec, Esther Keymolen, Tamar Sharon, Jan Smits,
Leslie Paul Thiele and the participants of the April 2019 ‘Governance of Emerging
Disruptive Technologies’workshop in Rotterdam for valuable comments on
earlier drafts. I thank Lisette Gotink for research assistance.
Disclosure statement
No potential conflict of interest was reported by the author(s).
Funding
The research for this paper was funded by the European Union’s Horizon 2020
research and innovation programme under grant agreement No 653626.
Notes on contributor
Prof. Dr. Bert-Jaap Koops is Professor of Regulation & Technology at the Tilburg
Institute for Law, Technology, and Society (TILT), the Netherlands. His main
research fields are cybercrime, cyber-investigation, privacy, and data protection.
28 B.-J. KOOPS