Mr Justice Eady :
The parties to the appeal
The Home Office, represented by Ms Steyn, appeals in accordance with s.59 of the Freedom of Information Act 2000 ("FOIA") against a decision of the Information Tribunal dated 30 January 2008. The Respondents are the British Union for the Abolition of Vivisection ("BUAV") and the Information Commissioner, who were represented respectively by Mr Alexander QC and Mr Choudhury. The primary challenge is to the Tribunal's construction of the terms of s.24 of the Animals (Scientific Procedures) Act 1986 ("ASPA"). It is necessary first, however, to set out the background against which the problem has arisen.
The use of animals in scientific procedures
The breeding and supply of animals for use in scientific procedures is governed by the provisions of ASPA, and the Home Office has important responsibilities for regulation in this context. Whenever it is proposed to use animals for scientific research, the circumstances are closely scrutinised. Applications need to be submitted and project licences may only be granted where certain conditions are fulfilled. It is necessary to demonstrate, for example, that the scientific procedures proposed will involve the use of the minimum number of animals; that the animals will have the lowest possible degree of neurophysiological sensitivity; that the experiments will cause the least pain, suffering, distress or lasting harm; and that they are most likely to produce satisfactory results. Where licences are granted, there is a code of practice laying down standards for animal welfare to which the relevant laboratory must adhere. There is a team of Home Office inspectors responsible for monitoring and inspecting the conduct of licensed laboratories.
It is clear from the evidence that those who seek licences from the Home Office for animal research will often be required to submit a great deal of detailed information beforehand which is sensitive or confidential for a variety of reasons. In particular, in order to satisfy the statutory requirements, it may be necessary for applicants to include material which is commercially sensitive, and/or potentially useful to competitors, and also details of locations and addresses which may be sensitive for security reasons.
When a licence is granted under this regime, the notification takes place by way of a covering letter to which the relevant licensee's application form is simply attached as a schedule (including, of course, all such confidential material as it contains).
The need for information to be made available
It is recognised by all concerned that legislators and members of the general public have a legitimate interest in knowing, if they wish, what is going on by way of animal research and the extent to which the regulatory functions of the Home Office are being properly discharged. The context of the present appeal is necessarily concerned with the extent to which information needs to be made available for that purpose.
There is evidence before the Court from Dr Richmond, head of the Animal Scientific Procedures Division at the Home Office, who was appointed in November 2003 and is the senior officialwith responsibility for policy and operational matters concerning ASPA. One of his tasks now is to deal with requests for information made under FOIA and giving responses which are as timely and accurate as possible. He has thus been grappling with these delicate issues since FOIA came into effect at the beginning of 2005.
The particular problem which forms the focus of this appeal is conveniently encapsulated in paragraph 9 of Dr Richmond's witness statement of 22 October 2007:
"The Department accepts that not all the information supplied on the form of application or appended as schedules to project licences was provided or held in confidence. Therefore we have sought various means to better inform the public debate about the licensing system and the programmes of work under ASPA."
Balancing freedom of information against a genuine need for confidentiality
It is recognised, conversely, that in keeping the public properly informed, and in responding to FOIA requests, it is necessary to have the means in place to safeguard material supplied by the relevant applicant or applicants insofar as it is genuinely confidential (either because it is commercially sensitive or because it consists of personal data). It is a difficult question to determine how this can best be achieved.
No doubt systems could be devised to achieve a fair balance between these important but conflicting interests, so as to ensure from the outset that it was clear to all concerned, and in particular to the applicants themselves, what information was to be treated on a confidential basis and what would be recognised as appropriate for public scrutiny if required. It is almost inevitable that such a system would, at any rate at the outset, be time-consuming and expensive to operate.
Dr Richmond attempts a reconciliation
Meanwhile, on the other hand, Dr Richmond and his department have been operating procedures which were adopted shortly after the FOIA came into effect under his guidance and supervision. In a letter of 14 December 2004, the Department wrote to the senior management at all sites where licensed animal experiments were taking place summarising how it was proposed that requests for information under FOIA would be addressed.
Shortly afterwards, on 21 December 2004, a further letter was written to all concerned, to explain the policy of encouraging those applying for licences, from then onwards, to prepare an abstract of their licensed work to be displayed on the Home Office website. In accordance with that policy, such abstracts have been published in respect of most newly granted project licences from December 2004 onwards. It is important to make clear, however, that these abstracts are drafted by the applicants themselves, that they are supplied voluntarily, and that they would normally be very brief (two to three pages) and very much shorter than the applications themselves. The evidence shows that the Home Office has been able to display abstracts, on this basis, in respect of 85% of the project licences granted in 2006 and that, in total, there are currently available around 1,000 published abstracts. Nevertheless, the adequacy of this system is under challenge.
It is important to have in mind that it has to be operated in a statutory context which forms the subject-matter of the present appeal and to which I shall shortly turn. It is thus not simply a question of arriving at an adequate flow of information on the basis of negotiation and compromise.
The factual background to the appeal
I must now summarise the factual background against which the present dispute arises. Within weeks of the new policy being announced, Dr Richmond received a letter from the Chief Executive of BUAV which contained a request under FOIA. Once the abstracts had been posted on the Home Office website in December 2004, in respect of nine project licences, the Chief Executive (Mr Sansolini) thought it appropriate to make a request seeking:
"the actual information contained in each of the following licences (using the titles given to the abstracts):
Relief from chronic pain by use of antidepressants;
Studying disorders of balance;
Metabolism and excretion studies for new candidate drugs; and
Genetically Modified Animals & Respiratory Diseases."
Dr Richmond decided to treat the letter as representing five separate requests under FOIA, acknowledging receipt on 18 January 2005. As he anticipated, the process of dealing with these requests proved labour-intensive and took some time. A letter was sent on 9 February 2005 to keep Mr Sansolini informed of progress meanwhile.
Dr Richmond was inevitably feeling his way and made a decision based on his earlier experience in supplying information in accordance with the 1997 Code of Practice on Access to Government Information (which he describes as "the forerunner of the FOIA"). He had found that the effect of providing documents which had been redacted (so as to conceal the information thought to be confidential) meant that the end product was often disjointed and difficult to follow. He therefore decided that it would be better to provide a narrative document which embodied all the information that it was thought appropriate to disclose. In the course of preparing such documents, steps were taken to consult those who had supplied the relevant information in support of their applications.
In due course a substantive response was sent to Mr Sansolini by letter dated 15 March 2005. There was an annexe which contained some further information relating to the five licences; that is to say, information which went beyond that contained in the published extracts. Dr Richmond had satisfied himself, as best he could, that it was legitimate to disclose this, in the sense that none of it was exempt information under FOIA. He explained that there were, on the other hand, a number of statutory exemptions which applied to the rest of the information requested, in the light of which it had been withheld.
The statutory context
For the purposes of the present appeal, the relevant exemptions arise through the interaction of two sets of statutory provisions. There are two sections of FOIA itself which require to be considered:
"41. Information provided in confidence
(1) Information is exempt information if –
(a) it was obtained by the public authority from any other person (including another public authority), and
(b) the disclosure of the information to the public (otherwise than under this Act) by the public authority holding it would constitute a breach of confidence actionable by that or any other person.
(2) The duty to confirm or deny does not arise if, or to the extent that, the confirmation or denial that would have to be given to comply with section 1(1)(a) would (apart from this Act) constitute an actionable breach of confidence.
44. Prohibitions on disclosure
(1) Information is exempt information if its disclosure (otherwise than under this Act) by the public authority holding it –
(a) is prohibited by or under any enactment,
(b) is incompatible with any Community obligation, or
(c) would constitute or be punishable as a contempt of court.
(2) The duty to confirm or deny does not arise if the confirmation or denial that would have to be given to comply with section 1(1)(a) would (apart from this Act) fall within any of paragraphs (a) to (c) of subsection (1)."
Alongside the FOIA regime, there falls to be considered an earlier statutory provision contained in s.24 of ASPA:
"24. Protection of confidential information
(1) A person is guilty of an offence if otherwise than for the purposes of discharging his functions under this Act he discloses any information which has been obtained by him in the exercise of those functions and which he knows or has reasonable grounds for believing to have been given in confidence.
(2) A person guilty of an offence under this section shall be liable –
(a) on conviction on indictment, to imprisonment for a term not exceeding two years or to a fine or to both;
(b) on summary conviction, to imprisonment for a term not exceeding six months or to a fine not exceeding the statutory maximum or to both."
There are thus two sets of provisions governing the protection of confidential information in this context (i.e. s.41 of FOIA and s.24 of ASPA). What is relevant for the current appeal is the prohibition on disclosure contained in s.44(1)(a) of FOIA, because Dr Richmond took the view that it would be a criminal offence for him to disclose any further information to BUAV, relating to the five licences in question, than that already revealed in his letter of 15 March 2005. He considered that it fell within the terms of s.24 of ASPA and was thus "prohibited by or under any enactment". No doubt he had the benefit of legal advice in arriving at his conclusion. A more general question has inevitably been brought into sharp focus, in the course of submissions, as to how far it is possible for the aims underlying these two statutory regimes (FOIA and ASPA) to be given effect compatibly with one another.
It is clear from the material before me that a positive decision was taken by the government to retain s.24 of ASPA alongside the provisions of FOIA, although other statutory restrictions were repealed to make way for a greater flow of information. Some light is thrown on this decision by remarks made on 1 July 2004 by Baroness Scotland in the House of Lords:
"Section 24, the so-called confidentiality clause, prohibits the disclosure by Home Office Ministers and officials of confidential information relating to the use of animals in scientific procedures other than in the discharge of their functions under the 1986 Act. It creates a criminal offence and provides a maximum punishment of two years' imprisonment and a fine for unauthorised disclosure of information. The Government undertook to review Section 24 in the context of our commitment to freedom of information.
The conflicting views expressed about Section 24 in the course of the review have shown it to be a complex and contentious issue which does not lend itself to a decision that will satisfy all of those with an interest. The House of Lords Select Committee on Animals in Scientific Procedures, as well as animal protection and anti-vivisection groups, have called for the repeal of Section 24. Those representing the views of many in the scientific community, whilst supporting other means of progressing openness and transparency about the use of animals in scientific research, have expressed significant concern about the potential impact of repealing Section 24 and have urged strongly that it be retained.
In the circumstances, we have concluded that Section 24 should be retained for the time being. We do not, however, rule out the future repeal of Section 24 and, therefore, propose to review the matter again in two years' time."
It would not appear, therefore, that the legislature was hoping to achieve any positive policy objective in retaining s.24; it was rather a matter of "wait and see". Meanwhile, it is for the civil servants, and where necessary the courts, to attempt to operate these two statutes, dating from different periods, and to reconcile them as far as possible.
If s.24 of ASPA were repealed, there would presumably remain in effect the other restriction to protect confidential information contained in s.41 of FOIA. For information to be "exempt" thereunder, its disclosure would have to "constitute a breach of confidence actionable by that or any other person". In the course of argument, it seemed that it was going to be an important question whether, for so long as s.24 of ASPA remains in effect, confidential information can be protected on a different and (apparently) more restrictive basis.
The procedural steps leading to this appeal
I must now summarise briefly the procedural history which has led to this appeal. BUAV sought to challenge Dr Richmond's interpretation and appealed to the Information Commissioner who, after a considerable delay, issued a decision notice on 12 June 2007 dismissing BUAV's appeal and agreeing with the Home Office's interpretation of the law. Because there was, on that view, an absolute exemption, the Commissioner did not go on to consider the other exemptions upon which the Home Office had placed reliance (i.e. under ss.21, 38, 40, 41 and 43 of FOIA).
The next step was for BUAV to go to the Tribunal, where a hearing was conducted on 17 and 18 December 2007. Mr Alexander's submissions on behalf of BUAV there found favour, and s.24 of ASPA was construed as requiring, for information to be classified as having been "given in confidence", that it be demonstrated that there would be an actionable breach of confidence if it were revealed in response to an FOIA request. It was thus given a similar meaning, in this respect, to that conveyed by s.41 of FOIA. In the result, the Tribunal directed that the Home Office reconsider its interpretation of s.44 of FOIA (although Dobbs J granted a stay of the Tribunal's directions on 29 February 2008, pending the outcome of this appeal).
The reasoning of the Tribunal emerges, perhaps, most clearly from paragraphs 16 and 17 of the decision:
"16. On this issue we prefer the BUAV arguments. We think that, even though section 24 ASPA does not make specific reference to the law of confidence, the use of the phrase 'given in confidence' means that the information in question was entitled to protection under that law – it means that it was given in circumstances where, because of the nature of the information, the circumstances of the disclosure and the harm likely to result from disclosure, the person receiving the licence application had a legally enforceable obligation to keep it confidential. The effect of the Home Office's argument would be that the threshold for criminal liability in this area would be lower than that for civil liability. That would be a remarkable outcome and we do not believe that it can be right.
17. The importing of the law of confidence in this way has the advantage that it provides a set of well established rules, based on case law, to be applied by a public authority when assessing information. We think that this is greatly preferable to the alternative, under which the test to be applied by the public authority would be very imprecise. The problems likely to be faced by a public authority in those circumstances are highlighted by the inconsistency we see between the interpretation which the Home Office has urged us to apply and the manner in which it has itself treated the BUAV request. If it were right that the only test to be applied was whether the information had been passed to the Home Office in circumstances that were capable of giving rise to an obligation of confidence, then it would not be necessary, or appropriate, for it to make any separation between disclosable and non-disclosable information. Yet that is what it has done in conceding that not all the information contained in the licence applications may be withheld. In releasing additional information, in the circumstances described in paragraph 4 above [as contained in Dr Richmond's letter of 15 March 2005], it has evidently applied criteria based on the nature of the information and not just the circumstances [in] which it was disclosed to it."
In referring to "a set of well established rules, based on case law", the Tribunal clearly had in mind the decision in Coco v A N Clark (Engineers) Ltd  RPC 41, 47-48, and what was described in paragraph 12 of the Tribunal's decision as "the following three part test":
"a. Does the information in question have the necessary quality of confidence?
b. If so, was it disclosed in circumstances that gave rise to an obligation to maintain its confidentiality?
c. Would its disclosure in breach of that obligation cause harm to the person who made the original, confidential disclosure?"
The Tribunal also thought it right to import the notion of "public interest" into s.24, so as to impose upon the Home Office civil servants (or, more strictly, the Secretary of State) an obligation to weigh up, in every case of prima facie confidentiality, whether there was an overriding public interest which would justify revealing the information nevertheless. The basis for this interpretation was the undoubted fact that in claims for breach of confidence the law recognises, in some cases, a defence of public interest: see e.g. Lion Laboratories v Evans  QB 526 and Att.-Gen. v Guardian Newspapers (No 2)  1 AC 109. It is right to point out, however, that there looms a daunting scenario (in hypothetical criminal proceedings) of having to investigate whether a defendant had not only (a) reasonable grounds to believe that information had originally been "given in confidence", but also (b) reasonable grounds to believe that there was a countervailing and overriding public interest in breaching the confidence.
The Tribunal's flawed interpretation of the law of confidence
The Tribunal rather proceeded on the assumption that "the law of confidence" was to be found only in the principles explained by Sir Robert Megarry in Coco v Clark. It assumed that this authority provided an exclusive definition such that, whenever the phrase "in confidence" was to be found in a statute, the legislature must be taken to have had those principles in mind. With respect, however, this does not seem to me to be necessarily the case. Much will depend on context.
It is clear, for example, that the law of confidence is not confined to the principles governing the circumstances in which an equitable duty of confidence will arise; nor to the specialist field of commercial secrets. An obligation of confidence can arise by reason of an agreement, express or implied, and presumably also by the imposition of a statutory duty. Nowadays, in addition, it is recognised that there is a distinction to be drawn between "old-fashioned breach of confidence" and the tort now characterised as "misuse of private information": see e.g. per Lord Nicholls in Campbell v MGN Ltd  2 AC 457 at  and the discussion by Buxton LJ in McKennitt v Ash  QB 73, at 80 et seq., under the heading "A taxonomy of the law of privacy and confidence".
Mr Alexander described Coco v Clark as being "then and now the leading authority on breach of confidence". But there would seem to be traps for the unwary in placing unqualified reliance upon the case without paying due regard to what Lord Nicholls had to say about it in Campbell v MGN Ltd in the section of his speech entitled "Breach of confidence: misuse of private information":
"13. The common law or, more precisely, courts of equity have long afforded protection to the wrongful use of private information by means of the cause of action which became known as breach of confidence. A breach of confidence was restrained as a form of unconscionable conduct, akin to a breach of trust. Today this nomenclature is misleading. The breach of confidence label harks back to the time when the cause of action was based on improper use of information disclosed by one person to another in confidence. To attract protection the information had to be of a confidential nature. But the gist of the cause of action was that information of this character had been disclosed by one person to another in circumstances "importing an obligation of confidence" even though no contract of non-disclosure existed: see the classic exposition by Megarry J in Coco v A N Clark (Engineers) Ltd  RPC 41, 47-48. The confidence referred to in the phrase "breach of confidence" was the confidence arising out of a confidential relationship.
14. This cause of action has now firmly shaken off the limiting constraint of the need for an initial confidential relationship. In doing so it has changed its nature. In this country this development was recognised clearly in the judgment of Lord Goff of Chieveley in Attorney General v Guardian Newspapers Ltd (No 2)  1 AC 109, 281. Now the law imposes a "duty of confidence" whenever a person receives information he knows or ought to know is fairly and reasonably to be regarded as confidential. Even this formulation is awkward. The continuing use of the phrase "duty of confidence" and the description of the information as "confidential" is not altogether comfortable. Information about an individual's private life would not, in ordinary usage, be called "confidential". The more natural description today is that such information is private. The essence of the tort is better encapsulated now as misuse of private information.
19. In applying this approach, and giving effect to the values protected by article 8, courts will often be aided by adopting the structure of article 8 in the same way as they now habitually apply the Strasbourg court's approach to article 10 when resolving questions concerning freedom of expression. Articles 8 and 10 call for a more explicit analysis of competing considerations than the three traditional requirements of the cause of action for breach of confidence identified in Coco v A N Clark (Engineers) Ltd  RPC 41."
Lord Woolf CJ had already observed in A v B plc  QB 195 at  that, because s.6 of the Human Rights Act 1998 required the court not to act in a way which is incompatible with a Convention right, the rights protected by Articles 8 and 10 had been absorbed into the long-established action for breach of confidence. As he there concluded:
"This involves giving a new strength and breadth to the action so that it accommodates the requirements of those articles."
The matter was also touched upon by Patten J in Murray v Express Newspapers Plc  EMLR 22 at :
"The incorporation of convention values into this branch of the law widens the focus of the cause of action to include private information which would never have been regarded as confidential by a court of equity in the days of Prince Albert v Strange (1849) 2 De G. & Sm. or even those of Coco v A N Clark (Engineers) Ltd  R.P.C. 41."
(The Court of Appeal heard an appeal from Patten J in March of this year, but the judgments are yet to be handed down.)
It is thus important to bear in mind, for the present case, the broad principle, stated by Buxton LJ in McKennitt at , that " … in order to find the rules of the English law of breach of confidence we now have to look in the jurisprudence of articles 8 and 10". The Tribunal did not address these developments at all and thus proceeded on an incomplete understanding of the present law.
It is also beyond question that some information, especially in the context of personal matters, may be treated as private, even though it is quite trivial in nature and not such as to have about it any inherent "quality of confidence": see e.g. Browne v Associated Newspapers Ltd  QB 103, 113-114 at , McKennitt v Ash (cited above) at  and the remarks of Lord Nicholls in Campbell v MGN Ltd (cited above) at . Thus, an obligation of confidentiality may sometimes arise in respect of such information merely because it is imparted as being confidential, either expressly or impliedly. Also, the law may imply an obligation on the basis that a communication has taken place in the context of an established relationship, which would itself give rise to such a duty.
The language and concepts used in Coco v Clark may still be apt in the context of commercial secrets and a duty of confidence owed in respect of them. Yet since this is not the only form of confidence, there is no warrant for supposing that the language used in s.24 of ASPA ("given in confidence") is limited to commercial secrets. It was, I think, accepted at the Bar that applicants for licences under the ASPA regime include private matters in their applications, apart from commercial secrets, which they may legitimately wish to keep confidential. If so, it would be inconsistent to argue that everything must be available for public inspection unless it can be restrictively defined as having about it the "quality of confidence". I would thus reject Mr Alexander's submission that "one cannot give 'in confidence' information which does not have the quality of confidence about it". It is too broadly stated.
Another way of putting the point would be to say that the law will afford protection, sometimes, where only the second of the Coco v Clark tests is satisfied; that is to say, the right to protection arises because it is clear to those concerned that the circumstances in which the information was imparted themselves give rise to a reasonable expectation of privacy. I would prefer, however, not to be tied to Coco v Clark where it simply has no application. (It was not even cited in the Court of Appeal in McKennitt, Browne or HRH Prince of Wales v Associated Newspapers Ltd  Ch 57.)
How to construe ASPA as the law of confidence now stands
It would seem to follow that it may be something of a distraction to attach too much significance to the distinction in wording between s.41 of FOIA and s.24 of ASPA, since in the light of the modern authorities there is no reason to suppose that even an "actionable" breach of confidence, where sued upon, must inevitably be founded on the formulation of Sir Robert Megarry.
One should try to give words contained in a statute their ordinary English meaning unless the context is such as to make clear that some specialist interpretation is required. All I need say in this context is that it does not seem to me that the phrase "given in confidence" must necessarily be given a specialist interpretation, to be defined solely by reference to Coco v Clark.
I think it fair to say, as counsel agreed, that s.24 of ASPA is by no means easy to interpret. In the words of Baroness Scotland, it is "a complex and contentious issue". There are a number of preliminary points to be made about it. First, it is a penal provision and should be construed accordingly.
Secondly, by contrast with the terms of s.41 of the FOIA, no express reference is made to an actionable breach of confidence. (Whether that, in the end, has any significance is another matter.)
Thirdly, nothing is said expressly about the "public interest" or as to who is responsible for deciding where it lies.
Fourthly, again by contrast with s.41 of the FOIA, this provision expressly contemplates a need to decide whether an individual "knows or has reasonable grounds for believing [information] to have been given in confidence".
Fifthly, s.24 of ASPA would appear to focus upon a particular point in time by reason of the past tense; that is to say, it is necessary to ask whether the hypothetical defendant in the criminal proceedings has a belief that the information was given in confidence. The statute does not require the hypothetical defendant to make an objective assessment of whether the information given, or parts of it, had a "quality of confidence"; in other words, he is not called upon to decide that information that was given in confidence is nonetheless not entitled to protection. One must concentrate, therefore, on the state of affairs arising at the time when the information was imparted. This is distinguishable from the situation contemplated in s.41 of FOIA, since that is focused upon whether or not there would be an actionable breach of a duty subsisting at the time when the relevant information would be revealed (by way of responding to a FOIA request).
Sixthly, the hypothetical defendant whose state of mind is being addressed in s.24 of ASPA must be taken to be a lay person (presumably a minister or civil servant working in the Home Office). It is surely questionable, especially in a penal context, whether Parliament intended that such a person should be taken to have knowledge of the circumstances in which an equitable duty of confidence arises and/or that criminal liability was to depend, in whole or in part, upon the reasonableness of an assessment as to where the public interest lay as to the disclosure, or otherwise, of a corpus of (largely technical) data.
Against this background, and in particular the rather unusual statutory context, it is now necessary to address the rival contentions of the parties. Mr Alexander seeks to uphold the decision of the Tribunal, whereas Ms Steyn (supported by Mr Choudhury on behalf of the Commissioner) challenges it for a number of reasons. She has drawn attention to some of the points I have noted above, and especially to the contrasts that can be drawn between this provision and s.41 of FOIA. More specifically, however, she argues that the Tribunal's construction would actually frustrate the object and purpose underlying the enactment of s.24 of ASPA.
Because Parliament considered it necessary to create a criminal offence (albeit only as a "longstop", as it was described by Viscount Davidson during the Parliamentary debate on 16 January 1986), it was clearly thought important at that time to afford protection to those who were required by the state to provide information to the government. Yet this was before so much importance was attached to freedom of information and open government. Until the complications brought about by the introduction of the 1997 Code of Practice, it would have been natural to assume that applicants supplying sensitive information for the purpose of obtaining a licence had a reasonable expectation that it would be treated in confidence. This blanket approach, however, was withdrawn in October 1998 (apparently following a challenge by way of judicial review). Thereafter, as the letter of 14 December 2004 explained, requests for disclosure of project licence information were approached case by case.
One difficulty about putting in place a structure which requires an ex post facto assessment to be made, of what should have been treated as confidential, is that the unfortunate applicant cannot be in a position to know what will be protected, and what will not, at the time the information is supplied to the Home Office.
A further difficulty, all the more acute because of the penal context, is that s.24 of ASPA would only provide the protection to which Parliament originally attached importance if the relevant civil servants were able to identify, straightforwardly, what information has been given in confidence and vice versa. Ms Steyn highlighted the irony that, applying its own Coco v Clark criteria, the Tribunal was not able, at the conclusion of a two-day hearing, to arrive at a decision on this important issue. The matter was sent back for further consideration by Home Office civil servants (inevitably through the spectacles of hindsight). Moreover, it was recognised that this would involve a considerable "burden of work". It is fair to record, however, that BUAV had suggested that this was the appropriate course to take, simply because the Home Office had greater familiarity with the subject-matter and also had access to the views of relevant licence holders. Even so, it seems most unlikely that in 1986 Parliament would have intended such a painstaking exercise (and one that is so uncertain as to its outcome) to be carried out in testing the "reasonable grounds for belief" for the purpose of establishing criminal guilt.
On the subject of the "public interest", Ms Steyn makes the point that when Parliament enacted s.24 of ASPA there would, inevitably, have been a weighing up of the public interest by the legislators, as a result of which their 1986 priorities found expression in this form of wording (according priority to confidentiality). It would seem an unreasonable interpretation to import, by implication, the need for a separate and independent assessment of the public interest to be made by civil servants (with the benefit of hindsight and with a different set of public interest priorities). Yet, as Mr Alexander responds, " … the public interest is itself an integral part of the law of confidence and does not need to be spelt out". As a general proposition, that is unexceptionable. It is again the context of a penal statute which makes it more difficult to apply.
Striking a note of judicial caution
It is necessary to have in mind, as I have already said, the fact that consideration has been given on a number of occasions to whether or not some of the pre-existing statutory bars on the disclosure of information should be removed in order to achieve compliance with the objectives of FOIA. These, for so long as they exist, have the effect of prevailing over the general right of access to information provided for in s.1 of FOIA itself. It is beyond doubt that s.24 of ASPA has been deliberately retained, notwithstanding the move to greater openness more generally. I should probably assume, therefore, that it was included to provide some additional, or at least parallel, protection to that contained in s.41 of FOIA, and judges need to be very wary of interpreting it so as to water down its effect. It may be, in reality, that there was no very clear-cut intention at all (as Baroness Scotland's comments might suggest), but that is not an appropriate assumption for a court to make when applying principles of statutory construction.
As Lord Lowry observed in C v DPP  1 AC 1, 28:
"(1) If the solution is doubtful, the judges should beware of imposing their own remedy.
(2) Caution should prevail if Parliament has rejected opportunities of clearing up a known difficulty or has legislated, while leaving the difficulty untouched.
(3) Disputed matters of social policy are less suitable areas for judicial intervention than purely legal problems.
(4) Fundamental legal doctrines should not be lightly set aside.
(5) Judges should not make a change unless they can achieve finality and certainty."
These were described by Lord Lowry as "some aids to navigation across an uncertainly charted sea", which he had derived from an analysis of the authorities. They are perhaps useful for me to remember in the present context. (I bear in mind that Lord Lowry's comments were made prior to the enactment of the Human Rights Act 1998 and that nowadays, in some circumstances, judges may be obliged to be more interventionist in the light of that development.)
Accordingly, a judge (or a tribunal) would need to think long and hard before deciding to step in (where legislators have hitherto "feared to tread") in order to introduce, by supposedly necessary implication, a whole set of criteria such as "public interest" and "quality of confidence" to which the legislators have made no reference even in debate.
Is there any help to be derived from Parliamentary materials?
It is sometimes interesting, and can be illuminating, to investigate Parliamentary debates in trying to understand the legislative purpose of an enactment, but it is clear that this should not be done unless the exercise yields a clear answer on a disputed interpretation: see e.g. R v Secretary of State for the Environment, Transport and the Regions, ex parte Spath Holme Ltd  2 AC 349, 391-392, per Lord Bingham.
I am quite satisfied that reference to Parliamentary material in the present case yields no such clear answer – let alone one that points to the Tribunal's interpretation. In concluding that the legislative history of s.24 of ASPA lent it support the Tribunal fell, in my judgment, into error. It follows that I should seek to avoid the same error. My construction, therefore, is based on the statutory wording, its context, the modern law of confidence and the arguments of counsel addressed to those matters – but without any reliance upon the content of Parliamentary debates.
My conclusion on the merits of the appeal
My conclusion is to uphold the Appellant's submissions to the effect that:
(1) the information sought by BUAV was obtained by the Home Office in the exercise of its ASPA functions;
(2) the relevant official (Dr Richmond) reasonably believed, as he has deposed, that it had been "given in confidence" at the time it was given;
(3) the disclosure sought from the Home Office would not be for the purpose of exercising its functions under ASPA;
(4) the terms of s.24 of ASPA, as it stands, mean that disclosure is prohibited and would constitute a criminal offence; and
(5) the FOIA exemption contained in s.44 applies to the information sought.
Accordingly, the appeal is allowed.
The remaining difficulties posed by the penal provisions of ASPA
None of this makes life any easier for Dr Richmond and his colleagues; nor for those who wish to use FOIA to achieve a fuller understanding of how ASPA regulation is working. There is obviously a tension between the terms of s.24 and the legislature's objective of greater public access to information underlying FOIA. They lie uncomfortably together and there is a need for legislative reconciliation. It is not for judges to square the circle.
It seems unworkable and unfair (as well as obviously very time-consuming) to leave the boundaries of confidentiality, and the weighing of public interest, to civil servants to work out subjectively by jobbing backwards to the time the particular application for a licence was made. What seems to have happened, when the additional material was released in March 2005, is that following consultation with the relevant applicants Dr Richmond decided that either it had not in fact been given in confidence, originally, or that there had been a subsequent waiver.
One of Mr Alexander's concerns is that on the Home Office's interpretation of s.24 of ASPA there could still be a finding of guilt even if the information had ceased to be regarded as confidential at some point. Because of the past tense, to which I have already referred, that may well be the correct interpretation. But it is highly theoretical. It is not conceivable that this "longstop" criminal sanction would be invoked when it could serve no purpose. It is not there for the protection of information which is no longer regarded as confidential.
It is unsatisfactory for those seeking information (such as BUAV), and inimical to the objectives underlying FOIA, to leave it to individual licence applicants to decide how much of their information is to be withheld from disclosure under FOIA. That would seem to be inherent in Dr Richmond's system of publishing "extracts", described at  above. On the other hand, it would be unfair to leave applicants in a state of uncertainty as to how much of the information they are required to give will be treated as confidential. This in turn may inhibit full and frank disclosure, and/or possibly the flow of such applications, and thus undermine the purposes of ASPA.
BUAV's proposals for the future
These considerations would appear to point to a need for matters to be sorted out at the time the application is made. There need to be clear criteria for identifying those categories of information in respect of which applicants should have a reasonable expectation of privacy/confidentiality. Fairness would require that, even if they cannot be allowed to define the limits for themselves, they should nonetheless have the right to be consulted and to make representations (preferably in the light of known criteria) as to where the boundary should be drawn.
Mr Alexander has made (if I may respectfully say so) a good start, in paragraph 69 of his skeleton argument, in trying to identify in general terms (and on the basis, obviously, of limited information) some categories of material in relation to animal experimentation which should normally not be regarded as attracting a cloak of confidentiality. His list was not intended to be exhaustive, but included the permissible purpose (by reference to s.5(3) of ASPA); duration of the project; background, objectives and potential benefits; justification, on the particular facts, for the use of primates, cats, dogs or equidae; the description of the plan of work; the index of procedures or protocols; housing conditions and environmental enrichment. He naturally recognises that these general categories would have to be subject to specific exceptions on some occasions, when genuinely sensitive matters have to be touched upon (usually commercial secrets). But it would surely be desirable to have available common and recognised criteria of the kind he suggests. Once these are worked out, and publicly available, the process of identifying confidential material at the application stage should become more routine and less time-consuming.
It would appear sensible, so that all those concerned know where they stand, to adopt as the starting point the presumption that the content of applications should be generally available, but to allow for confidential schedules to be attached. There are no doubt many who would agree with BUAV's case that "… as much as possible of the information needs to be publicly available in order to facilitate public, Parliamentary, and ultimately judicial, scrutiny of performance by the Secretary of State of her statutory duties". The problem is how to cater for legitimate exceptions without casting the net too widely. Once clear and acceptable criteria for permitting exclusion were established, it would presumably be relatively rare for interested observers (such as BUAV) to need to make challenges to the inclusion of material in the schedule. It would be advisable, no doubt, for something along the lines Mr Alexander suggests to be worked out and sanctioned by the legislature.
None of this is strictly for me to consider at all. I am only mentioning it in the present context to illustrate what a huge task is now involved, without such guidelines, and how burdensome it is to impose it on civil servants to work out ex post facto and on a case by case basis. That is particularly so, given the potential criminal sanctions in the background. If there gradually developed a system of confidential schedules, it would be relatively easy to impose a criminal sanction (if still thought necessary) specifically geared to unauthorised revelation of their contents. Everyone would know where they stood. These objectives may be desirable, and largely consistent with Parliament's policy on freedom of information, but I cannot see how this can be achieved consistently with maintaining the penal regime under s.24, at least in its present form.
Coping with s.24 of ASPA in the meantime
In the meantime, all that can be done by civil servants in Dr Richmond's position is to identify with hindsight how much of the information was originally supplied by the applicant in confidence; that is to say, according to either what was expressed to be confidential or what would have been understood (by a reasonable onlooker at the time) as having been given in confidence: see the language used by Lord Nicholls in Campbell v MGN Ltd at , quoted above.
In order to answer the question, it may be necessary to take a number of factors into account, such as communications between the applicant and the Home Office, as well as the surrounding circumstances at the time of disclosure. Naturally, it will be appropriate to consider the nature of the information itself, but (contrary to the Tribunal's view) that is not the only relevant factor to be assessed; nor can it be determinative.
On the wording of s.24 of ASPA, it would be reasonable sometimes to conclude that a stipulation for confidentiality on the part of the applicant, if it was accepted and went unchallenged at the Home Office, will have been sufficient to impose a corresponding duty of non-disclosure – even if the information concerned would not, when judged objectively, be characterised as a commercial secret or as having the "quality of confidence".
Whether this restriction is compatible with FOIA purposes, and whether it will stifle the free flow of information sought to be achieved, will be for legislators to decide in due course. Whenever s.24 is next considered, the ultimate conclusion may well be that it requires to be repealed or amended in the pursuit of freedom of information.