Rationality and Rhetoric in the Corporate World: The Corporate Annual Report as an Aristotelian Genre

This paper is part of a research programme into corporate annual reports. Reports do provide the information on the past performance, present state and future prospects which investors in listed companies require for the rational choices attributed to them. They also reveal the companies' responsiveness to the publics comprising the civil societies in which they are embedded. This effect requires more than strict rationality. To use Simon's distinction, the reports then entail both substantive and procedural rationalities. We argue that classical rhetoric and its recovery in the 'new rhetoric' yield useful approaches to the latter, and that annual reports comprise a genre in the rhetorical sense. We illustrate our case through generic features in the reports of the Australian-based multinational, Amcor. We suggest for future research that accounts of corporate functioning are incomplete unless they include the pre-structured interaction between companies and their publics which we have shown here through rhetoric.


Introduction
We aim in this paper to add to the conceptual apparatus in the study of corporate annual reports. These have long been a focus in nancial, organisational and managerial analysis, and it is widely agreed that they entail more than the 'information' required for investors either to make the rationally calculated and utility-maximising choices attributed to them under the model of 'economic man' or to exercise the rational control formalised in stock-market listing rules. There is no more agreement here, however, on how to allow for arational effects than there is in the study of the bounded rationality of corporate functioning more generally. Simon's call for attention to the issue remains very much to the point: 'Reasonable men' reach 'reasonable' conclusions in circumstances where they have no prospect of applying classical models of substantive rationality. We know only imperfectly how they do it. We know even less whether the procedures they use in place of the inapplicable models have any merit-although most of us would choose them in preference to drawing lots. The study of procedural rationality in circumstances where attention is scarce, where problems are immensely complex, and where crucial information is absent presents a host of challenging and fundamental research problems to anyone who is interested in the rational allocation of scarce resources. 1 We add to the study of procedural rationality by showing that annual reports ful l the functions of a 'genre'.
We take our sense of 'genre' from classical rhetoric and from its revival in the 'new rhetoric'. 2 Since the classical rhetoricians had debated how a rhetor could allow for the entanglement of formal rationality in the legal, political and ethico-moral issues in civil society, their problem was similar to that facing the writers of annual reports. Since exponents of the 'new rhetoric' have stressed the importance of the audience in any rhetorical encounter, they suggest attention to how the reports are read. From either side, the concept of 'genre' subsumes the interaction given in the writing and reading of annual reports We have a trans-disciplinary aim in adding to analyses of the rhetoric of business activity and analysis. 3 The 'new rhetoric' was a multidisciplinary event, and the range of disciplines where annual reports are studied is just as wide. They are the focus of a critical tradition in accounting. 4 They have been studied in administration, 5 in management, 6 in organisational theory, 7 and in environmental economics. 8 They have been a vehicle for advice in nancial analysis, 9 and in public relations. 10 Our account here is intended to be applicable across all of these disciplines. It is a problem-centred and theory-building study which requires no speci cally disciplinary foundation.
Between this introduction and a brief conclusion, we develop the paper in three sections. In the rst we give the rhetorical grounding of 'genre', and show its relevance to annual reports. To illustrate our argument, we then introduce the Australian-based multinational, Amcor, and show that elements in its reports from 1970 to the present can be matched with the forensic, deliberative and epideictic genres of classical rhetoric; the case of inherent con ict between capital and labour shows how Amcor resolves procedurally what is substantively irresolvable. We return to theory in the third section, to discuss the implications of a more general reading of the reports as rhetorical.

Classical Rhetoric and the Three Genres
'Rhetoric' is most familiar now in the pejorative sense of 'mere rhetoric', as the puff of vainglorious politicians. That usage suggests that the rational can be taken for granted, and that the factual is distinct from the way it is presented. In contrast, classical rhetoricians treated the rational and the factual as always arguable, as embedded in civic and cultural processes, and thus as never more than practically, consensually and provisionally closed. Rhetoric in this sense is immediately consonant with Simon's 'procedural rationality'.
One key issue in classical rhetoric was the arguability of contrary positions. Sophists such as Protagoras held that both sides of a dispute might be valid, that any strictly logical attempt to resolve them led to paradox, and that rationality then always entailed practical rationalisation. Plato famously attacked those claims in his defence of dialectic certainty. Holding that rhetoric required appeals to prejudice, he called it a panderer's knack, unworthy of the name of 'art' used by its apologists: 'it has no rational account to give of the nature of the various things which it offers. I refuse to give the title of art to anything irrational'. 11 Aristotle synthesised those themes. He favoured the Sophists in opening his tendentiously titled Art of Rhetoric, where he held that rhetoric was not inferior to but rather was 'the counterpart of dialectic', 12 and he accepted the claim that contrary positions could be argued. But he also allowed for Plato's critique by de ning rhetoric analytically: 'its function is not persuasion. It is rather the detection of the persuasive aspects of each matter'. 13 He thus treated rhetoric as a means of sustaining a focus on tensions between the rational and the arational. These were evident in the interaction between rhetor and audience, and this in turn was situationally variable. Aristotle discussed this effect under 'genre'.
When he adapted a typology of forensic, deliberative and epideictic/celebratory genres, he held that these three were logically necessary: the listener must be either a spectator or a judge, and, if a judge, one either of the past or the future. The judge, then, about the future, is the assembly member, the judge about the past is the juror, and the assessor of capacity is the spectator, so that there must needs be three types of rhetorical speech … 14 Forensic discourse was, narrowly, the oratory of the court, and more broadly any attack on or defence of particular actions. As a dissection of the past, it was always a vehicle for either the xing of blame or the allocating of praise. Deliberative speech was concerned speci cally with politics, and was applied more generally to any attempt to affect the course of affairs. Directed to the future, it typically involved a choice either to do or not to do something. Epideictic discourse referred to ceremonial or occasioned speeches, with the orator more intent on pleasing or inspiring an audience than on persuading it. Rather than being a matter of debate, the subject of an epideictic discourse allowed the crystallisation of consensus in the here and now, as in the enacting of national unity through speeches in honour of soldiers killed in action. Unlike the critically disruptive thrust in forensic and deliberative speeches, the epideictic was then conservative. It was practised by those who 'defend the traditional and accepted values … not the new and revolutionary values which stir up controversy and polemics'. 15 Given the intertwining of past, future and present in any event, this taxonomy of genres is more ideal-typical than descriptive. In practice, any discourse must be a combination of the three, with one or another genre emphasised at different points. Allowing for the expectations of various audiences under various circumstances, the taxonomy both describes and enacts the entanglement of the legal, the political/ economic and the ethical/social which was characteristic of civil society in Aristotle's time. Writers in the 'new rhetoric' have made some use of it, as when Gross held that a scienti c report 'is forensic because it reconstructs past science in a way most likely to support its claims; it is deliberative because it intends to direct future research; it is epideictic because it is a celebration of appropriate methods'. 16 More typically, however, they have de ned 'genre' in such general terms as 'typi ed rhetorical actions based in recurrent situations', 17 or as 'ready solutions to similar appearing problems'. 18 Once such recurrent situations and problems have been identi ed, any generic text might be studied 'much as an anthropologist sees a material artefact from an ancient civilization, as a product that has particular functions, that ts into a system of functions and other artefacts'. 19 Corporate annual reports comprise a genre in both those senses: they are discourses directed to the past, future and present of corporate activity; they are addressed to recurrent problems; and they are elements in the system of corporate functions. The typology of forensic, deliberative and epideictic genres then yields a practical rst step towards locating the effects of the reports in both that system and the more general functioning of civil society.
Trends in the analysis of annual reports suggest the usefulness of treating them as generic. We have already noted that they are studied in a variety of disciplines, and analysts certainly have reason to focus on them. Since they are 'the most publicized and visible document[s] produced by publicly owned companies', 20 since they 'communicate implicit beliefs about the organization and its relationships with the surrounding world', 21 and since they 'have the advantage of unobtrusive measurement in that they are written for purposes and to audiences different from [academic] analysts', 22 they give unique access to an organisation's embeddedness in civic society.
Analysts seem to have resisted treating the functioning of this embeddedness as generic. In discussing how annual reports might contribute to a 'corporate brand', for example, Ind criticised 'a tendency towards sameness in the tone of the reporting. There appears to be an innate conservatism of approach and a blandness to many reports'. 23 To some extent this is statutory. Since much of what a report should include is either legislated or set in stock-market listing rules, it would be startling were there not a 'certain sameness' in a company's reviews of its achievements, of its projected growth and of its current standing. Ind tacitly granted that when he held that reports 'should be delivering three core messages: historical performance, an insight into the company's future and an indication of management capability'. 24 Despite his explicit critique of 'conservatism' and his implicit denial of the generic, then, he still evoked the conservative effects of genre, for his 'core messages' entail forensic, deliberative and epideictic effects.
That speci c echo resonates too with the more general sense of genre as a cluster of typi ed responses to similar problems in recurrent situations, for the management of any company faces the annual problem of simultaneously satisfying contrary expectations. Their potential audience is obviously composite. While they ostensibly address their reports to shareholders, managers must allow for them being read by competitors, consumers, suppliers, regulators, pressure groups, the press, the market, trade union of cials, and present and future employees. The demands of each of these elements in the audience are often incompatible. That occurs, for example, in claims made for regulation and deregulation, in the tensions between different divisions within corporate networks, in disputes between workers and environmentalists, and in the con ict between capital and labour. Given those constraints, management faces a dif cult task in deriving the necessary appearance of unity. Their reports require a practical resolution of dif culties which, so far at least, have been found rationally irresolvable. The annual reports then meet the conditions which Aristotle and the Sophists ascribed to rhetorical situations in general: the need to crystallise consensus when opposed positions are both tenable.
So it makes good sense to treat the reports as generic. We now illustrate that argument by showing forensic, deliberative and epideictic effects in the annual reports of one company, Amcor, from 1970 to 1999.

Generic Effects in Amcor's Annual Reports
Amcor is well suited to a case study of allowance for contrary demands, for it is both typical and atypical of how Australian companies have responded to the shifts in global capital which have had a marked effect on Australian political economy. Throughout much of the twentieth century Australian industry was highly protected, the Australian workforce was highly unionised, and the Australian electorate was highly polarised. After peaking in the election of a (union-based) Labor government in 1972, after 23 years of (liberal-conservative) coalition rule, and in its dismissal from of ce three years later, con ict continued during the eight years of the coalition government rst elected in 1975, a period also marked by the economic stagnation which followed the world economic crisis of 1973. 25 An economic summit called in 1983 by the newly elected Labor government proved to be a watershed, for the corporatism set there allowed the rst steps towards the deregulation which has remained a feature of the Australian economy. The process was intensi ed on the re-election of a coalition government in 1996.
Under the name of A.P.M.-it was rebadged in 1986 as one mark of its increasingly global focus-Amcor was among Australia's largest listed companies in 1970, and has remained so amid the oating of the dollar, reduction of tariffs, large-scale privatisations and deregulation of the labour market which have transformed Australia. It has expanded from its base as the domestic market-leader in forestry, paper-making and packaging to become internationally signi cant in those industries. With levels of private share-ownership in Australia now among the world's highest, the number of its shareholders has shown a characteristic growth, more than tripling from 1970 to 1999. Given the overall concentration of capital in Australia, these owners are characteristically distributed. In the 1999 report, the 92.2% of them with 5,000 shares or less were shown as holding only 20.6% of issued shares. Conversely, the top 20 shareholders-all nancial institutions-held just over 50%. The company is just as interesting where it is atypical, for it straddles the recent shifts in civil society. It operates in the 'old economy' of manufacturing, and is well regarded for it. At the same time, it is susceptible to the 'new politics' of social movements, because of the direct and often unsightly exploitation of the environment which its industries require. The 'old economy' means that old con icts such as those between capital and labour remain in play; the 'new politics' adds new elements to the already diverse audience facing any large company.
So Amcor is well suited to our purposes. It is typical of major patterns in the Australian economy (itself typical of industrial economies world-wide) and its businesses require both a general responsiveness and a speci c resolution of the con icts typical of manufacturing industries. If our argument is justi ed, then forensic, deliberative and epideictic effects should be evident in the company's formally rational reports. So we need to show, rst, that Amcor does present a rational appearance, and then that this entails a forensic allocation of blame and praise for past results, a deliberative weighing of alternative futures, an epideictic celebration of present success, and the fusion of contrary elements in each of those in a presentation of consensual unity. As a speci c focus, we concentrate on how Amcor achieves consensus despite the con ict between capital and labour.

Amcor's Rationality
If Amcor's reports were taken at face value it would seem that the company is always under strictly rational control. It is obliged to present its accounts as mathematically rational, and it similarly uses the language of rationality in the more discursive components of the reports, where the board and management interpret the numbers showing pro t or loss, growth or decline, and success or failure. They routinely couch their interpretations in terms of the 'strategy', 'ef ciency', 'productivity', 'rationalisation' and 'restructuring' familiar from accounts of 'economic man'. Further, it was evident throughout the 30 reports we studied that the more uncertain the general environment became, the more the company stressed its rational responsiveness. As the old certainties of the Australian economy were removed, and as the company became more and more engaged in global competition and thus more and more vulnerable to shifts in the global economy, Amcor intensi ed its self-presentation as rational.
The company could obviously take this emphasis for granted, for it rarely made explicit the assumption of 'economic man' underpinning it. Only exceptional circumstances provoked direct statements, as when the board reacted in the 1976 report to what it saw as a politico-economic crisis in Australia by declaring that 'We advocate the principles of free enterprise where business and individuals can exercise initiative and gain rewards commensurate with their success'. That is an exception proving the rule, for in general Amcor took the grounds of substantive rationality as self-evident. The very language of that rationality, however, suggests the shakiness of its grounding. For example, once mathematically rational 'accounting' is translated into the narrative 'account' of an annual report it inevitably entails the nger-pointing of 'accountability'. Given that effect, the strictly rational must require the rhetorical counterpart which Aristotle suggested.
We could use any issue to show that process. In focusing on the relation between capital and labour we are following Amcor's own lead, for as it typically said in 1996, 'Our people are the key to Amcor's past, present and future success'. That conventional division of time points also to the division of forensic, epideictic and deliberative genres. We turn rst to the forensic genre.

The Forensic Genre
Classically, the forensic genre referred rst to the diagnosis of past events and then to the attribution of praise or blame for them; in annual reports, these points are covered in the required historical review of performance. Other writers have found that companies typically claim credit for their successes but attribute the blame for failure to external events, 26 and Amcor certainly ts that pattern. The obituaries for former directors and the encomia for retiring directors and executives which feature regularly in the reports, for example, leave no doubt as to who should take the credit for the company's success.
Here is one example, from the 1970 report: Your Directors record with regret the death of Sir Charles Booth, C.B.E., on 27th June, 1970. Sir Charles joined A.P.M. as a Director in 1944, He was appointed Managing Director in 1947 and served in that capacity until retirement from executive of ce at the end of 1958 when he was elected Chairman of Directors. He retired as Chairman in 1966 but continued as a Director of the Company until his death. It was under Sir Charles' leadership that A.P.M. embarked on its major expansion in the early post-war years. His faith in the Company and his vision were outstanding at all times. His long experience and wide knowledge of the industry contributed greatly to the prosperity of the Company and his advice and counsel will be sadly missed.
The tone was identical when the 1996 report included a note on the retirement of Stan Wallis, after 19 years as Managing Director: Mr Wallis joined the company in 1960 and was appointed Deputy Managing Director in 1975. Since then he has made an outstanding contribution to Amcor's continued growth and development in Australia and its expansion offshore. On behalf of all shareholders we thank him for his strong and imaginative leadership and for the important role he has played in establishing Amcor as a world-ranked packaging and paper manufacturer … More generally, management almost always praises the company's workforce. Thus, 'The year's [1968][1969] good results were achieved through the work and co-operation of staff and employees'; 'The progress of the Company during the year [1978][1979] results, in large measure, from the contribution and enthusiastic commitment of those who work for A.P.M'; and 'The excellent results achieved in 1988/89 re ect the skill and dedication of employees throughout the company'. Amcor then presents success as a collective achievement, under the guidance of its management.
It is quite otherwise with failure. Although Amcor experienced occasional setbacks, there is scant concession throughout the 30 reports that the board may have erred in setting policy, or management in executing it. Even in the 1990s, when the company underwent threatening reverses and when several major investments proved to be ill-chosen, there is scarcely any acknowledgement of board or managerial mistakes, and where there is, the point is tacit. The following passage from the 1999 report illustrates this effect: The improved results re ect the positive impact of major changes made throughout the company in the past few years. These have included changes in our organisational structure and senior management, rationalisation and restructuring of many of our businesses, the sale of a number of non-core or under-performing activities, a substantial expansion of our Australian ne papers business and a major and continuing cost reduction program.
This might imply that the board had previously been dilatory in setting the organisational structure and that senior management had been unsatisfactory, but that is as far as the company goes. Rather, Amcor routinely evokes four causes for disappointing results: economic conditions; governmental regulation; the intensity of competition in its markets; and, of most interest here, the same labour whom it ritually praises for success. During the 1970s and into the 1980s, A.P.M. routinely blamed strikes and industrial con ict when performance was unsatisfactory. This is from the 1971 report: Industrial disputes in the rst half of the nancial year detrimentally affected deliveries of papers and paperboards. Increases in productivity have been more than offset by substantial increases in costs during the year, notably in wages and in many other costs outside the Company's control, principally materials and transport. These factors were the main causes for the fall in pro t earned for the year from the manufacture of paper and paperboards.
Labour remained a cause for blame even under the corporatism of the 1980s. Thus the company noted in 1987 that: Negotiations between major packaging companies and the relevant unions regarding 'second tier' claims for additional remuneration bene ts have progressed satisfactorily except in Victoria where industrial disruption has caused major and unnecessary losses of production.
Although the company welcomed the attention at the time to deregulation of the labour market, it was impatient with the pace of the process. Thus it held in 1990, as one cause of a disappointing pro t, that 'Industry also is hampered by Australia's inability to carry out effective microeconomic reform on the waterfront and in our industrial relations, transport and telecommunications systems'. Even without the explicit reference to labour, microeconomic reform in Australia has often entailed union-breaking. To judge only by Amcor's reports, that process was successful, for after the earlier routine attention to unions and industrial con ict, the issues disappeared in the mid-to-late 1990s.
The company occasionally made explicit the tension in its blaming of sections among those whom it also praised, as when it wrote in 1976: In the earlier part of the nancial year our mills worked with little industrial disruption. However, in recent months there was an upsurge in industrial disputes. Over the full year the Company lost 40,000 tonnes of production through strikes and bans damaging both to A.P.M. and to our employees. Nevertheless, the majority of our staff and employees work hard and give loyalty and co-operation. Your Board places on record its appreciation of their contribution during another dif cult period.
Amcor then resolved the contradiction between praise and blame by reserving the latter for a disruptive minority. If that conventional scapegoating shows the entanglement of the forensic and the epideictic, then both are further enmeshed with the deliberative, for Amcor presented the loyalty and co-operation it praised as outcomes of its policy.

The Deliberative Genre
In classical rhetoric, deliberative speech referred speci cally to politics and more generally to any policy-making in civic affairs. It was entwined with the other two genres in that it entailed a response to diagnosis of the past and an attempt at consensus in the present. Amcor, of course, adopted policies on labour. As it said in 1977, 'We recognise for A.P.M. to remain ef cient the Company must have a loyal and well trained workforce'. It routinely included labour relations among its goals. In 1981, for example, it noted its intention to 'remain a good employer concerned with the safety, work satisfaction and overall welfare of employees', and it listed among its key objectives in 1995 its aim to 'maintain our signi cant commitment to mutually-bene cial employee relations through safe working conditions, training programs and recognition of the productivity and potential of our employees'. The taken-for-granted version of the past given by 'retain' and 'maintain' is worth noting here, and so too is the moral patina given to the mixture of 'ef ciency' and 'loyalty' by Amcor's use of 'must'.
Again, Amcor appears to take the assumptions behind that gloss as self-evident. There is a tone of restrained impatience when it invokes them explicitly, as when it stressed in 1972 that its: ability to provide good wages and working conditions is linked with cooperation by employees and Unions in achieving improved ef ciency and the elimination of unjusti ed strikes, and the acceptance of modernisation and appropriate crewing numbers.
In keeping with the asymmetry between praise and blame, the reports never mention strikes which may have been justi ed; modernisation and its associated rationalisation appear to be their own justi cation. The company also applied that sense of obviousness on a larger scale. It claimed in 1983, for example, amid the emerging corporatism of the newly elected Labor government, that: The continuing expectation of increasing incomes and more extensive welfare and community services can be justi ed only if we are able to achieve real increases in national productivity. We are hopeful that the Federal Government, armed with the spirit of consensus developed at the Economic Summit, will continue to lead the community to a better understanding of these principles.
Amcor reported a range of labour-related policies towards improvement in its own productivity.
One set of policies concerned the 'communication' by which it tried to instil in its workforce a morally loaded version of its substantive rationality. Thus the company claimed in 1978 that it: recognises the importance of effective communications and our policy is to develop a better understanding between management, employees and unions to encourage: -an increasing concern for, and pride in, the progress of the Company; -responsible attitudes towards production ef ciency and industrial relations; -recognition that A.P.M.'s success depends upon co-operation at all levels and between the various functions in the Company.
It said similarly in 1987: Considerable effort has been devoted to communicating to all A.P.M. employees our strategic plans and underlying objective of achieving international cost competitiveness. Our aim is to foster the acceptance and implementation of participative productivity improvement techniques. Local and company wide consultative processes have been established, with the support of relevant unions, to reassess training needs and rede ne job requirements, quality and reliability standards for import replacement and export development and attitudes to manufacturing activities. This action aims to recognise the often hidden potential that already exists in the organisation and to use these resources to achieve improved results.
Amcor did more than talk about this 'communication'. It also gave material expression to its calls for better 'understanding', for 'mutually bene cial' labour relations and for 'participative productivity improvement techniques'.
In the mid-1980s the company developed a policy for employees to buy shares at a discount. Giving the workforce 'a direct stake in the company's future growth and prosperity' (1988), and offering it an 'incentive to strive for improved results and to share in the bene ts of Amcor's success' (1989), the offer was widely taken up. At a peak in 1996, more than 14,000 of the company's 25,000 employees were listed as shareholders. Since introduction of the plan coincided with the declining salience of unions, the alignment of interests it represented appeared to mark the company's success in resolving tensions between capital and labour.
That success, however, was not complete, and Amcor continued to urge policies to shift the remaining tensions in its favour. Thus it treated the high unemployment during the recession of the early 1990s as more an opportunity than a problem. It noted in 1991, for example, that: There will not be a better opportunity than during the current economic recession for management of Australian industry to reach a common understanding with the workforce on ways to reduce costs and improve ef ciency and productivity. Amcor has made impressive progress in the past few years in this regard, but if our businesses are to remain internationally competitive there must be reasonable incentive for us to continue to invest and expand in Australia.
Since it is clear enough on whose terms the 'reasonable' would be decided, this is a tacit concession of continuing con ict.
Although the reference to 'common understanding' then might have the edge of an offer which cannot be refused, the appeal to consensus also suggests the interweaving of the epideictic throughout this deliberative policy-making.

The Epideictic Genre
Referring to ritual discourse, the epideictic genre is immediately relevant to annual reports. Orators in that mode were less intent on persuading an audience to a particular position than on pleasing or inspiring it, as they crystallised consensus through the celebration of success or the commemoration of tragedy. Annual reports, of course, are typically devoted to the success required for corporate survival. Amcor's claim in 1996 that it 'has a strong record of growth and has consolidated its position as one of the world's leading packaging and paper companies' is a characteristic gesture in that direction. As already suggested in the forensic praise of employees and the deliberative policy towards a 'common understanding', inclusion of the workforce is crucial to this epideictic celebration.
We noted too that the appeal to unity in the epideictic genre is conservative in its results, in contrast to the potentially radical thrust of diagnosis or policy-making. That effect is suggested in the 'our people' which Amcor commonly uses in the reports to describe its employees. The phrase evokes a feudal sense of belonging and of mutual obligation rather than the self-interested and calculating rationality pervading the forensic and deliberative moments of the reports. Although Amcor rarely refers explicitly to 'corporate culture', with all the arationality which that entails, it does implicitly invoke a sense of collectivity beyond that of 'economic man'.
We have already suggested one means by which it does so, its use of scapegoating to resolve the tension between forensic praise and forensic blame of its workforce. That is a routine device in the reports. Here is another example, from the 1977 report: Our relationships with most unions are good; however, small groups both within A.P.M. and outside the Company are bent on disruption to the detriment of the workforce as a whole and some maintenance unions are using strikes and other work limitations to press their claims for pay increases and other bene ts outside the Government's indexation guidelines … Since the rhetorical antithesis which the company uses here always suggests the polarity of good and evil, its identi cation of a 'them' also identi es a uni ed and right-thinking 'us'. Isolation of a small minority allows the celebration of such putatively shared values as the 'commitment', 'dedication', 'enthusiasm', 'loyalty' and 'team-work' which Amcor regularly invokes.
It reinforces that effect by a means which will be evident in the passages we have quoted, its routine use of the rst person plural. When the chairman, the managing director or divisional managers use 'we', 'us' and 'our', they sometimes have a speci c referent in the board or in groups within the company. But rst person plural pronouns have the advantage of imprecision, and their use allows more to be implied than stated. Phrases like 'our team of highly skilled people' (1981), 'our most important resource' (1989) or 'our management and employees' (1992), for example, gloss the differences between owners, management and labour. They also invoke the readers' lived senses of collectivity beyond the company. Here is a not atypical example, an extract from the 1983 report which we have used above: The continuing expectation of increasing incomes and more extensive welfare and community services can be justi ed only if we are able to achieve real increases in national productivity. We are hopeful that the Federal Government, armed with the spirit of consensus developed at the Economic Summit, will continue to lead the community to a better understanding of these principles.
The rst 'we' denotes 'we Australians'; the second refers narrowly to 'we, the board'. By eliciting the broad sense of collectivity and then by identifying itself with it, the board naturalises what it can expect from its workforce. To derive unity from disunity, that is, it transmutes the rational to what is obviously reasonable on the basis of national traditions.

Annual Reports as Generic
We have shown that Amcor uses its annual reports to communicate a forensic allocation of blame and praise for past results, a deliberative planning for the future, an epideictic celebration of present success, and the glossing of contrary elements in each of these in an appearance of consensual unity. Since those effects are evident in the one issue of the relation of capital and labour, we are con dent that they are more generally applicable. But we should also note three limits in our treatment of annual reports as generic.
First of all, the forensic, deliberative and epideictic genres of classical rhetoric give no more than a convenient framework. They are separable analytically, but only so long as the point of analysis is the interaction between their effects. That is why we have stressed throughout our discussion that the genres are mutually entwined.
Secondly, the apparent consensus which follows from the practical resolution of con ict in that entanglement is never more than provisional. Thus the appearance of unity which Amcor derived from its 'participative productivity improvement techniques' and from its aim of 'mutually bene cial employee relations' was vulnerable, as the 1998 report indicates: substantial abnormal losses were incurred during the year, mainly re ecting costs of plant closures and rationalisations of poorly performing businesses. Regrettably, this has involved job losses, but as a result Amcor is now more ef cient and competitive than ever before … The uni ed and arational 'dedication', 'loyalty' and 'commitment' which the company claimed to value were clearly more negotiable than it allowed. It implicitly expected retrenched employee-shareholders to welcome as owners what they might have found devastating as workers. Many of Amcor's employees faced those con icting rationalities, as shown in Table 1. We cannot tell how many of the 6,000 employees who left the share register were among the 5,500 retrenched in the cause of ef ciency, but it is fair to assume that the two groups had many common members. It is equally striking that 8,000 remained as shareholders despite this demonstration of how precarious their positions were.
Our third limitation, then, concerns the need to study annual reports in conjunction with studies of the meanings of share-ownership, for Amcor's remaining employee-shareholders have clearly made sense of their situation. Although we have shown the generic constraints which the company faces, and although we have shown some of the rhetorical moves it makes in allowing for them, we have relied only on our own reading for our account.
These limitations are important, but they do not lessen the signi cance of our study. Our demonstration of forensic, deliberative and epideictic moments in Amcor's reports, and of its generic allowance for contrary demands, warrants a more general reading of annual reports as rhetorical. We turn now to some theoretical implications of that approach.

Rhetoric and Annual Reports
The rhetorical working of corporate communication has implications well beyond the scope of this paper. We touch here on only two issues: the long-standing tensions between formal and informal rationalities; and the 'control' associated with rationality in much organisational analysis. To develop those points requires that we return brie y to classical rhetoric, and to the 'new rhetoric' of its adaptation.
As we stressed when we introduced classical rhetoric, disputes over rationality have a long history, and our generic study of annual reports is in that sense nothing new. Rhetoric might even have originated in the study of business operations, for tradition has it that although it had long been taught orally it was rst formalised by Corax and Tisias amid the political upheaval and property disputes following the death of Hiero of Syracuse in 466 BC. 27 A legend about them highlights the tension between formal and rhetorical rationalities which Plato attacked in his disputes with the Sophists and which is still a feature of rationality in practice. Impressed by Corax's success in the assembly and courts of the new republic, Tisias approached him for tuition. The two agreed that Tisias would pay for his instruction once he had proved its worth by winning his rst case. As soon as the lessons were completed, Corax sued for payment, arguing that: 'If I win, I win; if I lose, I also win, by the terms of the contract'. But Tisias responded: 'If I win, I win; and if I lose, then I too also win, by the terms of the contract'. The judges took the only sensible decision, and drove them both from the court. The story may be apocryphal, but its point remains. There is a permanent tension between apodeictic appeals to reason and epideictic appeals to consensus in formal and informal rationalities.
That tension might almost de ne the social sciences. Some disciplines and subdisciplines are based on the assumption of a rationally calculating and utility-maximising 'economic man', and others on the observation that the analytical bene ts deriving from that assumption entail an oversimpli cation of rationality in practice. Simon's call for attention to 'procedural rationality' in situations where attention is scarce, where problems are complicated, and where information is lacking is but one among many attempts to restore that pre-empted complexity. Even if analysts who have allowed for arationality have often met the fate of Corax and Tisias, the issues raised in classical rhetoric remain permanently relevant in the study of economic activity.
The 'new rhetoric' on which we have also drawn, however, is not just a return to classical themes, for the interaction between speakers/writers and their audiences stressed in it entails a crucial shift of emphasis. Where the rhetor was the focus of classical rhetoric, the new rhetoric is de ned by attention to listeners or readers. When incompatible statements are equally reasonable, 'the appeal to reason must be identi ed not as an appeal to a single truth but instead as an appeal for the adherence of an audience'. 28 Unless an attempt to persuade is adapted to its listeners' or readers' expectations, it is like the question-begging of formal logic, in that the rhetor would presume the agreement at issue. It follows that 'the image of the powerful orator playing masterfully with the emotions of the helpless crowd is a myth. … [I]f orators can control crowds, it is only because crowds control orators'. 29 That stress on the audience ts the new rhetoric to the age of mass democracy, mass communication, and mass involvement in corporate activity.
If the rising levels of private share-ownership throughout the post-industrialising world do mark a shift from 'managerial capitalism' to 'investor capitalism', 30 then it becomes more pressing than ever to study the effect on corporate operations of investors' expectations. The new rhetoric yields a new approach to the old tension between ownership and control, for when a company ful ls forensic, deliberative and epideictic expectations it is yielding to a form of control from below. There is then a certain 'metaphysical pathos' in rationalised accounts of rationalised managerial control. 31 Shareholders are the ostensible audience for annual reports and since we have shown generic effect in the reports we can infer the constraints imposed on rms by what they expect those shareholders expect. Even if the t between what a company does and what it says it does is only rough-and most analysts of annual reports agree that it is generally better than that-it is still moot as to who is controlling whom. Just as orators can control crowds only because crowds control orators, we suggest that a company controls its shareholders only if its shareholders control it. Thus the blends of substantive ef ciency and procedural loyalty which we found in Amcor's reports are at the same time constrained by readers' expectations of logical and ethical consistency and enabled by those readers' more general experience of and allowance for inconsistency. Amcor's surviving employee-shareholders show that effect clearly enough. To over-stress the rational is to miss that generic interaction.
That effect should not be over-stressed. Managers do occupy strategic positions, and not all shareholders/shareholdings are equal. Amcor's 20 largest shareholders own more than 50% of its shares, and it cannot be doubted that they disproportionately exercise substantive control. But even when that imbalance is granted, procedural/rhetorical control still remains in play. Since the block shareholders are nancial institutions which in turn must satisfy their own shareholders and investors, the effects of interaction are simply shifted one step. Since shareholders are not the only audience for annual reports, those effects are more general still. Any company must also take into account the often incompatible demands made by its potential readers. As Amcor's deft allowance for the con ict between capital and labour shows, what is substantively impossible can be resolved procedurally/rhetorically, within the constraints of a fuzzy control from below.
Classical rhetoric is then a useful supplement to analyses in which rationality is taken as unproblematic. Such a supplement is permanently necessary, for although 'economic man' has been queried ever since he made his rationally calculating and utility-maximising debut, he still strides through much of the literature. The emphasis on the audience in the new rhetoric gives an extra edge to that approach, in re-opening access to the issue of 'control'. Those are among the bene ts which follow from treating corporate annual reports as a genre.

Conclusion
We began this study of annual reports with Simon's remark on the under-studied problem of how 'reasonable men' reach 'reasonable' conclusions when they have no prospect of applying classical models of substantive rationality. We have argued that both classical rhetoric and the 'new rhetoric' are useful approaches to that dif culty, and have illustrated their potential by focusing on the narrow question of how one company, Amcor, reconciles rational ef ciency with its stated aim of fostering a loyal and committed workforce. The generic effects which we found in Amcor's reports showed that Aristotle's stress on the epideictic as the counterpart of the apodeictic yields an interactive approach to substantive and procedural rationalities: substantive rationality is a necessary but not suf cient condition of reasonableness; the rationality of economic man is among the arational values enacted procedurally. For a company to present a reasonable responsiveness to its shareholders then requires that it draw on both forms of rationality.
Since we have argued for Amcor's success in that regard on the basis of its nancial success we need to add another note of caution here. Amcor used a rising share price in 1999 as evidence of its success in restructuring, but its share price is languishing as we write this, for the company has been neglected amid the booming demand for shares in the new technology companies. It is too early to assess claims that the old economy has passed; Amcor's successful divestment of its paper-making activities, which also occurred as we were writing this article, suggests it is alive and well. But we are con dent that however Amcor balances forensic diagnosis, deliberative policy-making and epideictic celebration in its response to that falling share price and that divestment in its annual report for 2000, the account will be as plausible as the pride it took at the market's response to it in 1999. That is generically required, and Amcor has proven adept at meeting demands which are simultaneously defensive, informational and ceremonial.
When attention is scarce, when problems are complex, and when information is absent, both rms and their shareholders make what sense they can of their uncertain embedding in processes beyond their control and of their uncertain relation with each other. They do so by drawing on what can be taken for granted and by attending only selectively to contradictions in that taken-for-grantedness. When embeddedness entails legal, politico-economic and socio-ethical elements, the rhetoric developed to account for their intertwining in civil society is a promising, and perhaps necessary, focus for research. In later papers we aim to bring that focus to such issues raised in annual reports as corporate governance, corporate reputation, corporate social and environmental responsibility, and corporate strategy. The over-rationalised variants on stakeholder theory and agency theory commonly used in these elds require supplementing with the attention to the arational which we have treated here through 'genre'.