QJM Advance Access published online on June 29, 2009
QJM, doi:10.1093/qjmed/hcp084
When I use a word ... Declarative titles
‘Can saying make it so?’
J. L. Austin, ‘How to do things with words’
William James Lectures, 1955 (OUP, 1962)
This piece has an indicative title—it tells you what it is about. I could have given it a declarative title, one that declares that something is so, such as ‘Declarative titles are deceptive’.
The Indo-European root KLA, to call or cry out, shout, resound, has spawned a wide range of words, such as claim, acclaim, declaim, exclaim, proclaim and reclaim; clamour; clarinet and clarion; class and classify. Of the seven different entries for the noun ‘clam’ in the Oxford English Dictionary, the one relevant here is defined as ‘the crash caused by ringing two or more bells of a peal together’; the bivalve that implies the opposite, i.e. silently clamming up, is from another root entirely. Clear and chiaroscuro, clarify and clairvoyant are from clarus, a Latin derivative meaning loud or sonorous (of noise); bright, shining, clear (of light); distinct, unambiguous, clear (of words); and illustrious, famous, notorious (of people). Clarus gives declarare, to make known, announce, indicate, show and (of words) to express a meaning, signify.
In computing, declarative refers to programming languages that specify properties rather than procedures, distinct from imperative languages. A program in a declarative language ‘explicitly states what properties the desired result is required to exhibit but does not state how the desired result is to be obtained ... (what rather than how)’.1 It is the same in science in general—declarative titles reduce science to a product rather than a process.2
Declarative titles have also been called assertive titles.2 The definition of ‘assertive’ in the OED starts quietly but ends pejoratively: ‘characterized by assertion; declaratory, affirmative; positive, dogmatic’. And ‘self-assertive’ is defined as ‘insisting on one's own rights or claims; egoistic; assuming’. Assertive titles have attitude.
Declarative and assertive titles are related to ‘performative’ sentences, first identified in his 1955 William James Lectures3 by John Langshaw Austin, a philosopher of language, ranked with Wittgenstein, and also the author of ‘Sense and Sensibilia’, the title being taken from ... well, Aristotle actually. In a performative sentence a thing is done, actually done, by being declared or asserted: ‘I marry you’, ‘I crown you king’, ‘[I bid] three no-trumps’, ‘I name this ship the QE2’, ‘I promise you’, ‘I bet you a pound’, ‘you are hereby warned ... ’, ‘I take this woman’ (to be my awful wedded wife, as Dylan Thomas put it), ‘I advise you [not to use declarative titles]’. In identifying this type of sentence, Austin overturned the philosophical notion, already contradicted by the grammatical evidence, that a statement can only be used to describe something that is true or false. Jonathan Miller has said that Austin's idea was the most important thing, intellectually, in his life: ‘I suddenly became dazzled and fascinated by the idea of doing things with words.’4
Declarative titles do something—they state something that may or may not be so, and they are formed performatively, since they imply ‘I assert that [declarative titles are deceptive]’.
Declarative titles in science journals hardly existed before 1970. Rosner reported that of 2582 titles published in the Proceedings of the National Academy of Sciences from 1960 to 1969, none was declarative; in 1970, 0.7% (3/429) were; by 1980 the figure had increased to 25% and after 1986 it was 
45%.2 The worst offenders were molecular biologists, which Rosner attributed to the declarative chapter titles in James Watson's influential textbook, The Molecular Biology of the Gene (1965). But the main factor, he thought, was the increasing pressure on all scientists to be goal oriented, leading to what he called the ‘title wars’.
However, in the last 15 years, declarative titles have found a new sponsor—evidence-based medicine. Using declarative titles was one of the declared procedures of the journal Evidence-Based Medicine, when it was launched in 1995,5 and in 2003, The Journal of Clinical Epidemiology started not only to publish declarative titles but also to commission them, claiming that such titles would be more informative.6 ‘Indicative’ titles, they said ‘give the purpose of a study’, and ‘declarative’ (or ‘informative’) titles ‘give the conclusion’. Although, as one declarative title has put it, ‘The evidence provided by a single trial is less reliable than its statistical analysis suggests’.7
Declarative titles generally use the past tense when describing the results of a single investigation (‘Aspirin relieved headache’) and the present tense when stating an accepted finding (‘Aspirin relieves headache’). However, this makes probabilistic findings appear definitive. Aspirin may have been significantly better than placebo in relieving headache. But that does not mean that it relieved everyone's headache. After enough studies we may accept that ‘aspirin relieves headaches’, but we still recognize that it is not a universal truth, despite the declarative title. You might accurately say ‘Aspirin often cures headaches’, but then we would have to agree on the meaning of ‘often’.8
The most insidious form of declarative title is the one that uses neither past nor present tense, but tells you the result as an apparently indicative statement. Consider ‘Symptomatic benefit from eradicating Helicobacter pylori infection in patients with nonulcer dyspepsia’.9 This implies ‘We assert that there is/was ...’. We do not know whether this means ‘eradicating Helicobacter pylori relieves symptoms’ or ‘eradicating Helicobacter pylori relieved symptoms’. We do not even know if it is probabilistically true in either case. For, as readers of the journal in which this title appeared eventually discovered, the outcome was actually uncertain.10
Defending the use of declarative titles, Richard Smith irrelevantly suggested that ‘ "Freddie Starr ate my hamster" is one of the great titles of our time. "Freddie Starr and my hamster: a personal account" would have long been forgotten.’11 Failing to distinguish definitive and probabilistic assertions, he did not meet the criticism that, as Goodman put it, ‘rejected hypotheses live on in such titles’.12 Goodman had reviewed 24 reports of clinical interventions published in 1996 with ‘prevents’ in their titles—the intervention had not been preventive at all in at least eight cases. Furthermore, Smith had to admit that he had failed to find solid evidence on even the readability of declarative titles, and contented himself with reporting that journalists ‘believed strongly’ in using them.
At best, declarative titles mislead; at worst they may enshrine a falsehood as a permanent truth. When they hide the probabilistic nature of even what evidence-based medicine regards as the best forms of evidence—randomized controlled trials and systematic reviews of such trials—declarative titles divert attention from those individuals who do not fit the supposedly universal paradigm. The perennial problem of particularizing from generalities remains.
If I wanted to find out if spironolactone would normalize the serum potassium concentration in unselected patients with hypokalaemia due to potassium depletion, I might carry out a randomized placebo-controlled trial, and would probably find that spironolactone produces significant benefit. A later meta-analysis might be headed ‘Spironolactone increases serum potassium in hypokalaemia’. Others would use spironolactone for this purpose in regular practice. One day, however, one would encounter a patient with Liddle's syndrome (a channelopathy that affects renal epithelial sodium channels), as I have13; her doctor could not understand why her hypokalaemia did not respond to spironolactone. But he should have used amiloride instead. Because doctors think that spironolactone always increases the serum potassium, and because it is more effective than amiloride in the types of cases that they most commonly encounter,14 they are found wanting when the exception occurs. This is a rare example, but it well illustrates the point.
Finally, consider some other possible apparently impeccable declarative assertions: ‘digoxin slows the ventricular rate in atrial fibrillation’ [but not in thyrotoxicosis]; ‘co-amoxiclav cures pneumonia’ [but not if it is due to MRSA]; ‘triptans are indicated in migraine’ [but not basilar or familial migraine]. Or ‘antiarrhythmic drugs are antiarrhythmic’; well yes they are, except, that is, when they are paradoxically proarrhythmic.15
References
1. Illingworth V, Glaser EL, Pyle IC. Dictionary of Computing. In: 3rd edn (1990) Oxford: Oxford University Press. 123–4.
2. Rosner JL. Reflections of science as a product. Nature (1990) 345:108.[CrossRef][Medline]
3. Austin JL. How to Do Things with Words. (1962) Oxford: Clarendon Press.
4. Dickson EJ. Telling stories. The books that change our lives. In: The Times(Books Section). 14 April 2001: 15.
5. Anonymous. Purpose and procedure (abbreviated). In: Evid Based Med (1995/6) 1–2.
6. McGowan J. Declarative titles. Accessed 23 April 2009. [http://www.chla-absc.ca/2005/Presentations/0601/McGowan_CHLA2005.pdf].
7. Borm GF, Lemmers O, Fransen J, Donders R. The evidence provided by a single trial is less reliable than its statistical analysis suggests. J Clin Epidemiol (2009) 62:711–15.[CrossRef][Web of Science][Medline]
8. Aronson JK. Sometimes, never. Br Med J (2006) 333:445.
9. McColl K, Murray L, El-Omar E, Dickson A, El-Nujumi A, Wirz A, et al. Symptomatic benefit from eradicating Helicobacter pylori infection in patients with nonulcer dyspepsia. N Engl J Med (1998) 339:1869–74.
10. Gray JAM. The unreliability of some knowledge hallmarks. In: In: Evidence-based Healthcare. How to Make Health Policy and Management Decisions, 2nd edn, section 5.10.4 (2001) Edinburgh: Churchill Livingstone.
11. Smith R. Informative titles in the BMJ. BMJ (2000) 320:915.
12. Goodman NW. Survey of active verbs in the titles of clinical trial reports. BMJ (2000) 320:914–15.
13. Aronson JK. Rational prescribing, appropriate prescribing. Br J Clin Pharmacol (2004) 57:229–30.[CrossRef][Web of Science][Medline]
14. Kohvakka A, Eisalo A, Manninen V. Maintenance of potassium balance during diuretic therapy. Acta Med Scand (1979) 205:319–24.[Web of Science][Medline]
15. Hauben M, Aronson JK. Paradoxical reactions: under-recognized adverse effects of drugs. Drug Saf (2006) 29:970.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||