Risk,Uncertainty and the British Atlantic Merchant Ship as a Technology for Profit,1600–1800

Humans, uniquely, produce artefacts of extraordinary complexity and are able to do so only by the successful development of networks of trust, cooperation and coordination. ¹

This article rests on the following assumption: if we do not adequately understand the risks and uncertainties facing shipowners and operators, then we will more than likely misunderstand their technological choices. Sending a ship to sea, to transport a cargo from one place to another in the hopes of making a profit, is an archetype of the risky venture. What I will focus on here is what my work has taught me over the past several years: that technological continuity and change in the ordinary British North Atlantic merchant ship, during the period of imperial establishment and growth, was primarily driven by the desire to mitigate the risks and uncertainties inherent in the use of that technology. Continuity and change in ship technology adapted accordingly. Rather than use ‘risk’ and ‘uncertainty’ interchangeably, however, or lump them together without defining them, I will here consider the distinction between them that economist Frank Knight insisted on in his influential 1921 Risk, Uncertainty and Profit, and attempt to determine what use that distinction might have in the examination of developments in seventeenth- and eighteenth-century British Atlantic merchant ship technology. ² As participants in this industry learned more about what they were doing, and about the environments, both natural and human, in which they were doing it, they were able to turn uncertainties into known risks – up to a point.

Risk versus uncertainty

Knight wrote that the key to the problem of profit could be found ‘in the notion of risk or uncertainty and the ambiguities concealed therein’. ³ He argued that ‘risk’ is a ‘measurable uncertainty’ and ‘uncertainty’ is ‘non-quantitive’. ⁴ Knight's idea immediately calls to mind marine insurance. Insurance actuaries attempt to quantify, and to predict, the chances of various hazards occurring to the assets they insure. The more confident they are in their ability to do so, the more willing they will be to write policies to protect those assets, and the more comfortable they will be lowering the premiums they charge for those policies, which in turn will make their ‘products’ more attractive to buyers, increasing the pool of premiums they collect and retain to pay claims and make a profit. The rise of actuarial insurance is a historical phenomenon of the second half of the eighteenth century in the British Atlantic, and it did indeed increase the availability, and consumption, of marine insurance. ⁵ By the 1750s, the maritime commercial community had a century and a half of experience with navigating the Atlantic Basin. The weather patterns and major currents, and their seasonal variations, were well understood. Charts were more accurate and more widely available. Navigational training was more widely disseminated. ⁶ Shipwrights, and at least rudimentary repair facilities, were widespread throughout the British Atlantic world. ⁷

Insurers also wrote into their policies stipulations and exclusions that were designed to protect the underwriters from bad decisions on the part of the insured. They wrote exclusions for sailing into certain conditions. They required sailing in convoy in circumstances they deemed to carry a high likelihood of predation. They began to insist on professional evaluations of the condition of the vessels they insured – hence Lloyd's Register, with its classification system. Knight writes that ‘the assumption of “risk” of error in one's own judgment, inherent in the making of a responsible decision, is a phenomenon of a different character from the assumption of “risk” in the insurance sense’. ⁸ Perhaps we can consider the premiums that insurers charged to be based on the other risks then – the risks Knight is calling risks ‘in the insurance sense’, which arise from sources beyond the judgment of the insured. So, we can see that here, indeed, we have two different things we are calling ‘risk’.

Risk is ‘a known chance’, while ‘true Uncertainty’ cannot be so ‘known’, and that ‘true Uncertainty’ is, for Knight, causally connected to profit. ⁹ Such uncertainty may be reduced – in fact, as Knight argues, it can be converted into risk – but it can never be eliminated entirely; if it could, we would have a system of ‘perfect competition’, and that would eliminate the opportunity for profit. ‘Chief among the simplifications of reality prerequisite to the achievement of perfect competition is … the assumption of practical omniscience on the part of every member of the competitive system’. ¹⁰ ‘Perfect competition’ does not exist, then, because ‘practical omniscience’ does not exist.

What does exist, however, is an increase in knowledge with experience. Participants in a certain industry – in our case, merchant shipping in the eighteenth-century British Atlantic – added to their collective knowledge of the tactics and strategies called for to better avoid the hazards presented to them. By employing successful tactics and strategies over time, through trial and error, they learned enough about those hazards that, to an increasing extent, they could convert those hazards from uncertainties to risks, to use Knight's concept. An important example would be Atlantic weather patterns. The first European crews sailing deep into the Atlantic Basin, and reaching the American coasts and islands, did not know what to expect from the weather. They did not know the average frequencies of storms or their intensities, trajectories and seasonality. They were unaware of the prevailing currents and winds in the western and southern sectors. Any underwriter asked to write an insurance policy for such a venture would have had almost no information on which to base any assessment of the chances for success or failure, and thus of the likelihood of a large claim. ‘Insurance’, however, encompasses more than a written policy promising to compensate owners for losses. The early ships themselves were insurance; small and stout and armed, the small galleons sent out by the pioneers of colonization were well suited to send into the unknown, with no expectation of support afield. ¹¹ They would be far less suitable for use on well-established shuttle trade routes with support facilities and human networks at both ends, and capable naval protection at sea. Other types would, under such circumstances, be far more profitable to the private firms competing for business that replaced the royal monopoly companies financing the early expeditions – companies set up to bear far more Knightian uncertainty than their successors would.

Uncertainty, for Knight, was more likely to manifest itself in ‘single instances’ than in groups of events that could lend themselves to statistical probability. ¹² Through repeated experience with a hazardous undertaking like ocean sailing, the humans engaged in that experience could take advantage of one of our core skills as a species: pattern recognition. The Atlantic weather patterns became predictable – not entirely so, of course (‘practical omniscience’, as noted, does not exist) – but to a degree that, in Knight's construct, moved those weather patterns from uncertainties to risks. The sea, however, retained plenty of uncertainties to throw at ships and their crews. On 26 June 1766, HM Schooner St. Lawrence was at anchor at Cape Breton, Nova Scotia. A bolt of lightning struck her mast and travelled down to her powder magazine. She blew up. ¹³ Random accidents were going to happen, and their randomness meant that making decisions to lower the chance of their occurrence was difficult to impossible. To a certain extent, the industry as a whole would have to bear the burden of them – and ‘bear’ is indeed the verb Knight employs. Knight argues that the consequences of uncertainties are diffused among a larger group of specialized actors, equipped to bear the burden of them and willing to spread out the penalties as they occur. These people are specialized because they possess the knowledge, experience, skill and assets best suited to cope with these uncertainties.

So, ‘uncertainties’ become ‘risks’ by becoming somewhat regular, groupable patterns with some degree of predictability. Knight writes that ‘the chief practical significance of knowledge is control’ – specifically, ‘control of the future’, which is ‘closely interrelated’ with ‘increased power of prediction’. ¹⁴ The ‘control’ made possible by ‘knowledge’ is not control of the hazard, of course, but control of the human response to the existence of that hazard and the possibility that it will interfere with a venture in pursuit of profit. Both control of the future and the increased power of prediction, Knight writes, ‘are closely identified with the general progress of civilization, the improvement of technology and the increase of knowledge’. ¹⁵ The contribution to the ‘increase of knowledge’ seems straightforward enough, but the ‘progress of civilization’ and the ‘improvement of technology’ are loaded, problematic concepts for historians of technology. ¹⁶ They also lend support to Jonathan Levy's characterization of Knight as a modernist – someone whose thought fit into the narrative of ‘modernity as a disenchanted universe, of cold instrumental rationality, legal and quantitative generalizability’. ¹⁷ ‘Progress’ and ‘improvement’ were core assumptions of modernism. Since Knight's time, we have been far more sceptical of them.

Continuity and innovation in ship technology

As for ‘progress’, I will limit my comments to technology, and thus I can address the notion of technological ‘improvement’ at the same time. These are, of course, value judgments; they depend on subjective assumptions of the sort that are so often tacitly assumed to be held in common across time and space. In order to defend the judgment that a technological adaptation – whether that adaptation was a choice for continuity or change – is an ‘improvement’ and thus constitutes ‘progress’, we need to show evidence that the adaptation made it easier for the people who made it to achieve whatever objective motivated them to do so.

General surveys of technological history, and even more specific surveys of ship technology, can convey the impression that ship design, construction and rigging during the period 1600–1800 was relatively static. ¹⁸ That impression would be reinforced by reading the economic history of Douglass North and his students James Shepherd and Gary Walton, for whom the only technological aspect of merchant shipping of any import was the presence or absence of defensive armament. ¹⁹ It is true that changes in the merchant ship were more modest and incremental than they had been in the periods immediately prior to 1600 and, especially, shortly after 1800. ²⁰ However, continuity only represents stagnation in a technology if it is holding back – serving as a drag on – the industry it is supposed to serve. If it is not, then continuity, like change, can represent a technological adaptation to changing conditions. This is an especially important lesson in technological history for our society, obsessed as it is with innovation – hyper-innovation, in fact, in which digital technology has become, colloquially at least, synonymous with ‘technology’ itself. There are far more important technological continuities in our society than many of us realize, as David Edgerton has pointed out; it seems to me that understanding continuity and change in earlier technologies can help us overcome what might otherwise be a particularly myopic grasp of technology and society in our world. ²¹

The ship itself carried some basic characteristics forward through the period, and those carried their own inherent set of hazards. ²² Since ships were built by fastening together hundreds of pieces of wood, the dangers of leaking, rot and fire are obvious. Less so are the dangers posed by loss of structural rigidity, which could result from leaking, rot or fire, of course, but could also arise from the design and construction incorporating insufficient built-in structural reinforcement for the operating conditions of the Atlantic. Building in sufficient structural rigidity dictated, to a certain extent, the designs of hulls, and thus the sailing characteristics, capacity, stability and rigging options of those hulls.

Some of that structural rigidity depended on the support of the water in which the ship floated. An advantage of the full, bluff, rounded bow sections of these vessels – a trait that persisted strongly throughout the period – was that it offered good buoyancy forward, supporting the front of the hull and lifting the bow up over waves rather than allowing the vessel to bury itself in them, transmitting a shock through the structure and reducing the forward momentum so important to a wind-powered vessel. Shipwrights paid a great deal of attention to preventing the deformation of hulls in the water over time due to hogging (drooping at the ends relative to the middle) or sagging (drooping at the middle relative to the ends). Both were highly detrimental to sailing characteristics, and both placed great structural stress on the vessel, ensuring chronic leaking and, eventually, the compromise of structural members. With experience, shipwrights learned construction techniques that minimized the likelihood of such compromise.

One price of that full bow shape was reduced maximum speed through the water, and the issue of speed is perhaps the easiest entrée into the technological choices made in the early modern British Atlantic merchant fleet. While it might seem irresistibly desirable to employ ships that can get from point of origin to destination as quickly as possible, in most cases, that benefit was not worth its cost to the merchant shippers who owned these vessels. First, to design and build a hull for maximum speed through the water, the shipwright must sacrifice capacity for size, as a hull with sharper lines, and thus less internal volume for cargo, will generally go faster. The usual way we think of this is in terms of the block coefficient, which sounds inaccessibly technical for a general scholarly audience but is actually easy to envision. ²³ Imagine a rectangular block of wood. Now imagine the shape of a ship's hull inside that block of wood, the exact same length, width and depth of the block (see Figure 1). The fuller the hull, the more of the block's volume the hull takes up. The sharper the hull, the less of the block's volume it takes up. The fuller hull will carry more and the sharper hull will go faster, all other things being equal. Merchant shippers always chose the fuller hull because it allowed their vessels to carry the most cargo for their size. That raises the question: Why not build a larger hull with sharper lines so that you could carry the same amount of cargo as the smaller, fuller ship but carry it faster? This question is even more compelling when we consider that longer hulls are intrinsically faster than shorter hulls, thanks to a principle of hydrodynamics. Eventually, in the nineteenth and twentieth centuries, this would be done. For the early modern British Atlantic, however, the answer to that question is the desire to mitigate the potential cost of one or more uncertainties turning out unfavourably.

OPEN IN VIEWER

Figure 1.

Block coefficient of a Ship's Hull.

Building a larger vessel was exponentially more costly than building a smaller one, as the volume of a hull increases exponentially as its length increases, all other proportions remaining the same. Shipwrights charged their clients by the builder's ton, a measured unit of volume. ²⁴ A larger hull also cost more to fill with cargo, and the only way the owners of that cargo could realize a profit on it was to sell it all. Otherwise, they would have incurred less cost and risk shipping a smaller cargo. The markets of the early modern Atlantic were growing, but they were not big enough to offset this risk. They were also volatile, largely agricultural and, especially given that communication of market conditions was only as fast as the fastest ship, unpredictable. A modest fluctuation in the price of a commodity could make the difference between making money and losing money. Larger ships also cost more to rig and equip. More importantly, however, they cost more to man, all other things being equal. Skilled maritime labour was so valuable, and the supply so limited, that the cost of labour pressured for technological change in merchant ships if they grew in size. In other words, all other things could not be equal if ships were to grow in size.

And yet, was it not necessary that they grew in size, given the demographic and economic growth of the early modern British Atlantic in these two centuries? Would not a resistance to increasing the size of merchant ships cause a drag on growth in a developing maritime economy? To some extent, yes, but uncertainties demanded an alternative to significant increases in the size of merchant ships. They demanded a dramatic increase in the size of the merchant fleet, while keeping size increases in individual ships modest and incremental.

We know to look for conditions specific to the Atlantic merchant fleet to explain the relatively modest size of its ships, even late in the period, when we consider the size of large warships and the ships of the East India Company. In both cases, the pressure to contain costs was not nearly as strong. The Royal Navy, of course, was not seeking profit. Different priorities dictated its approach to manning. The East India Company was operating in a different environment, both commercially and physically.

Cost control

So far, the pressure to contain labour costs has proven to be the best explanation for the aggregate of evidence related to technological change in the British Atlantic merchant ship. ²⁵ Changes in rigs, sail plans and steering systems all permitted somewhat larger vessels to be safely manned by the same or an only slightly higher number of men. No other possible explanation for these changes holds up as well as a causal explanation for any of them, though I should note that more interdisciplinary research is called for here. I am specifically calling for twin avenues of research on that theory: first, archival work on comparative building, rigging and manning costs for vessel types available as options for the same service – such as the sloop and the schooner, or the two-masted brig (or snow) and the three-masted ship of the same size – and, second, trials aboard appropriate operatingperiod replicas, primarily focused on comparative labour requirements. ²⁶

I will sum up what seems to be the case. First, the total sail area of ocean-going square-rigged ships was divided into more and smaller sails as time passed and ships grew somewhat in size. At a time when mechanical advantage on deck was strictly limited, this meant that the ship's total sail area could increase, to power its increased size, without making the sails too powerful for the crew to handle. In turn, a more divided sail plan helped make possible the increase in two-masted as opposed to three-masted square rigs for midsize ocean-going merchant vessels, a trend that increased significantly beginning around 1750. ²⁷ Eliminating an entire mast, with all of its sails and rigging, much of which had to be worked aloft, saved construction and rigging costs and allowed fewer men to work the rig. ²⁸ We also see more fore-and-aft sails on eighteenth-century vessels compared to their predecessors, whether combined with square sails in more thoroughly hybrid rigs than we see in earlier square-rigged ships, or as the primary driving sails on smaller vessels, particularly schooners and sloops. ²⁹ Fore-and-aft sails did not require the crew to go aloft to work them, and their rigging was usually simpler, though they had some performance drawbacks vis-à-vis square sails. ³⁰ Finally, after a long period of experimentation, wheel steering was installed on more and more midsize and larger vessels in the early eighteenth century, and offered more mechanical advantage as well as visibility from the helm than tiller steering on those classes of vessel. ³¹ In each case, the causal explanation best supported by all of the evidence is the desire to achieve modest increases in size, and thus capacity, not speed, without simultaneously increasing labour demand.

Shipowners could not compromise speed too much, however. The most burdensome vessels, most of them Dutch, operated in the Baltic Sea and North Sea, where the security environment placed less pressure on merchant vessels than in the Atlantic. ³² Not only was the North Atlantic much larger and subject to greater natural violence; it was also typically subject to greater human violence, as the European maritime powers fought each other and preyed on each other's merchant shipping, and free agents, whether attached to a state by a piece of paper called a letter of marque or a privateer's commission or operating independently as pirates, also posed a real threat to the merchant ship, its crew and its owners. The greatest cost that Atlantic merchant shipowners could not control, and had to shoulder throughout the period, was the cost of security. That central thrust of North's, and Shepherd and Walton's, argument seems to hold up well.

Cost versus security

Armament was probably a larger component of that cost than concessions to speed, given that the latter were also desirable for contending with not just the human, but also the natural, demands of the Atlantic operating environment. While we see a decrease in the arming of Atlantic merchant ships in the second half of the eighteenth century, with the increased effectiveness of the Royal Navy in the protection of shipping and the related growth in the marine insurance market, it never came close to disappearing, and merchants thus bore the costs of arming their vessels. ³³ Those costs went far beyond the cost of the weapons themselves, though that was significant. Carriage guns placed great strain on the structural members of the ship, requiring stronger and thus more expensive construction than unarmed vessels. That construction also made the vessel heavier, which meant that it needed more power to make the same speed through the water as a lighter vessel. More power meant more sail area, and thus more labour demands, in addition to extra cost for rigging and rigging maintenance. Armed vessels had to carry barrels of powder and racks of shot, both of which were heavy, took up considerable space and were potentially dangerous – especially the powder. It is an old argument that the high-risk (or high-uncertainty, as Knight might prefer) security environment of the Atlantic was a drag on the productivity of the British Atlantic shipping industry, and that the drag manifested itself in the technology of the ship itself, primarily in terms of armament. No evidence that I am aware of from the past 50 years contravenes that argument. The uncertainty presented by that environment might translate into high profit for a group of owners who made a good decision and got lucky, but might result in serious loss for another who made the same decision and did not get lucky.

Nick Burningham and I present evidence that two of the most popular rigs in the eighteenth-century British Atlantic merchant fleet were adopted only after they had been adopted by the Royal Navy, despite the fact that antecedents of these rigs had existed for some time. ³⁴ This, along with other evidence, suggests that merchant vessels copied elements of the appearance of naval vessels not only for prestige, but also for security. From a distance, a merchant vessel that resembles a naval vessel is less likely to be attacked by a predator. Security considerations also account for the presence of fake gun ports and even fake guns on merchant ships. ³⁵

The idea that merchant ships followed the Royal Navy's lead would also suggest that the reduction of castling – the high platforms built up at the bow and stern of seventeenth-century vessels, carried over from the preceding period – over the course of the eighteenth century in merchant ships followed the same trend in naval vessels. That reduction is seen primarily as a sort of reverse naval arms race; as ship-to-ship combat transitioned from grappling and boarding to the firing of cannon broadsides from some distance, the need for high fighting platforms for soldiers decreased, allowing shipwrights to forego the weight, expense, windage and stability challenges imposed by castling, which had completely disappeared by the second half of the eighteenth century. ³⁶ So, again, a security concern had held certain costs high, and the elimination of that security concern allowed those costs to come down – in this case, literally.

Mitigating risk and uncertainty

While owners might make technological choices to hold down their costs, thus increasing their chances of profit, ships were still so expensive to own and operate that most were fractionally owned. The early period of British Atlantic trade was dominated by the chartered monopoly companies, which were created to mitigate the otherwise overwhelming uncertainties of striking out into so much unknown, but the Virginia Company and Massachusetts Bay Company were no longer needed once the maritime economy of the Eastern Seaboard took off after the mid 1600s. The Hudson's Bay Company, trading in the far north, and the Royal African Company, transporting enslaved Africans, remained in those specialties, the latter with increasing private competition. For the most part, however, the British Atlantic shipping industry had outgrown its need for such organizations now that it operated with better-known, and even partially quantifiable, risks.

To mitigate risk and uncertainty, owners had to employ trusted shipwrights and, frequently, trusted representatives as liaisons to the shipyard. Owners had opinions about how they wanted their ships built, but they deferred to the shipwright for important matters of design and rigging, sometimes not even specifying the rig of the vessel at all but leaving that up to the shipwright. A reliance on such expertise, facilitated by a trusted expert agent of some sort, helped reduce the risk of being defrauded or of unwittingly purchasing an unsuitable or unfit vessel – one at a decided disadvantage when faced with the unavoidable risks and uncertainties of the Atlantic, where the lure of profit and the reality of peril were inseparable; in fact, to Knight, the former could not exist without the latter.

The shipwright, in turn, put his business reputation up for public scrutiny every time he built and launched a vessel. He had to balance pleasing his client with preserving his reputation, and this meant that shipwrights sometimes refused to accommodate client requests that, in their judgment, would compromise the vessel's integrity or seaworthiness. ³⁷ Shipwrights did experiment, but they had to be circumspect about it and refrain from engaging in untried innovations. The world's best mathematical-scientific minds were still working out the principles of hydrodynamics, and they did not succeed in doing so in the eighteenth century. This meant that the primary practical advantage of modern ‘naval architecture’ – a higher degree of predictive power – was unavailable; predictability was based on precedent. ³⁸

While this may seem at first glance to be a recipe for technological stagnation, such was not the actual result in this period. These shipwrights were building vessels that served their clients’ needs, at a cost their clients could afford to pay. There was no incentive for drastic innovation, which was far more likely to prove detrimental than beneficial. The price for a failed innovation in an ocean-going ship was likely to be painfully high; as Richard Unger puts it, ‘error had a very high cost’. ³⁹ That fact certainly encouraged a conservative approach to the craft, and that approach served the industry well. It allowed for a long period of incremental adjustments to the basic technology, preserving its proven benefits while adapting it enough to prevent it from becoming an impediment to the industry's ability to adapt.

Our best evidence suggests that the most important of these incremental adaptations to ships was the increasing of capacity for size, offering owners the potential for greater cargo-carrying profit while limiting building, labour, tax and insurance costs vis-à-vis a longer vessel. ⁴⁰ This was complementary to the modest increases in outright size noted earlier. It resulted from the incorporation of Dutch design elements into British (and French) hulls, which was noted as early as Ralph Davis's Rise of the English Shipping Industry and is supported by much more recent archaeological work. ⁴¹ While such adaptation would not make a vessel faster, speed was not a driving priority, as explained earlier, and, with a modest increase in size and thus waterline length, a slight decrease in speed due to a fuller hull might have been offset. Some laboratory experiments with ship-design software might be instructive here.

Another side benefit to a fuller hull, especially a fuller after section, would be in sailing performance, especially on the wind (upwind). As I explore in my book based on the schooner Sultana, the combination of a full bow section and a more tapering stern (rear) section contributed to an undesirable effect called griping, when the vessel was heeled (leaned over) by the wind. The effect was to encourage the ship to turn into the wind, which would have to be counteracted by adjustment to the sail plan, in the absence of which the crew would have to use excessive pressure on the helm (rudder) to resist the turn-in, leading to strain on the rudder, drag from water turbulence and thus reduction in speed. These are considerations we need to test if we are to understand fully the ramifications of an important technological development that we are already confident took place. ⁴² The ‘cod's head and mackerel's tail’ design was debated in the late eighteenth century, and it still has not been fully investigated in modern maritime history.

External mitigating factors

Finally, we should consider two important and related responses to hazards external to the ship, both of which increased in importance beginning in the mid eighteenth century. These were convoys and insurance. ⁴³ They are related because, once the Royal Navy had established its pre-eminence in the Atlantic Basin in the Seven Years War, it was better able to escort convoys effectively, which in turn encouraged merchant owners and masters to join convoys, despite their drawbacks. Those drawbacks frequently included significant delays in sailing while convoys were assembled and because convoys could only sail as fast as their slowest ship. Delays could cost the individual merchant ship, or the entire group, favourable market conditions at their destinations. An entire group of vessels arriving at a destination at the same time could also glut a local market or leave some of the vessels without a return cargo. Nevertheless, in the wars of the eighteenth century, convoys were good protection against predators – much better than they had been earlier, when they sometimes presented irresistible targets to enemy cruisers. Policies commonly included provisions for the payment of ransom, allowing owners to instruct masters to surrender to a predator rather than engage in combat, as the vessel would be allowed to proceed to its destination rather than seized. Also encouraging participation was the fact that eighteenth-century insurance policies, which were generally affordable, required sailing in convoy during periods of conflict or the serious threat of it. It was the combination of convoys and insurance that accounts for some reduction in defensive armament in the second half of the 1700s, permitting owners the reduction in costs associated with that.

As for piracy – maritime predation conducted without state sanction – its importance and the success of imperial states in suppressing it are still matters of debate, but the increasing power of the Royal Navy, and its ability to project that power westward and southward, acted as an effective counter, if not an outright solution, to all forms of predation on the merchant commerce of the Empire, and it seems clear that piracy did not comprise as significant a component of that threat in the later decades of the eighteenth century as it had earlier, especially on the high seas. ⁴⁴

Conclusion

The design, construction and rigging of these ships was conservative – evolutionary rather than revolutionary technology. ⁴⁵ Nevertheless, it was adapted, incrementally but significantly, to meet the needs of the shipping industry for economic and demographic growth while continuing to offer a robust response to hazards in a business and physical environment whose dangers remained real and ever-present. The greatest impediment to the growth of the shipping industry was not the physical environment of the Atlantic Ocean; while the hazards of the sea were significant and beyond human control, they could be known to a large and useful extent – and, by the eighteenth century, they were known and predictable to an extent that actuarial insurance against them was a viable business. The most formidable impediment to the growth of the shipping industry remained the threat of human violence, which was never far away for long in this period of Atlantic history, and kept maritime interests reliant on defensive armament, ships and crews that could bear it, and both naval and financial asset protection.

By the late eighteenth century, then, much of what had been unknown when the British established their Atlantic Empire was understood. Political and commercial organizations and strategies were in place to mitigate hazards. The skills of the shipwright and the navigator were better developed, and better informed by the conditions their ships would be subjected to in service. Yet, with each ship on each voyage, those invested in that venture still faced what Knight calls uncertainty, as opposed to risk, and what John Kay and Mervyn King prefer to call ‘radical uncertainty’, as opposed to ‘resolvable uncertainty’. ⁴⁶ Radical uncertainty refers to hazards whose chance of occurrence remains unknown. It lies beyond the limits of what actuaries can compute with any confidence. In our world, Kay and King argue, too many of those in charge of business ventures and economic policy mistakenly believe that radical uncertainty does not exist, and that they possess models and calculations capable of converting radical uncertainty into resolvable uncertainty. No eighteenth-century British Atlantic shipwright, shipowner, master or sailor believed anything of the sort. None would have discounted the inescapable power of luck or the unknowable will of God. We might assume that, absent what Kay and King take pains to point out is false confidence, these people would have been hampered in their efforts by some sense of helplessness. That they were not is usefully explored by Thomas Doerflinger in his aptly titled A Vigorous Spirit of Enterprise, a study of the maritime merchants of Philadelphia. ⁴⁷ That spirit would have been no surprise to Knight, who

believed that it was radical uncertainty that created profit opportunities for entrepreneurs and that it was their skill and luck in navigating radical uncertainty which drove technical and economic progress. Keynes, too, wrote that without ‘animal spirits’ and ‘spontaneous optimism’, ‘enterprise will fade and die’. ⁴⁸

These people did not fool themselves into thinking that they knew more than they did about the hazards they faced. Instead, with the ships they built, paid for and sailed, and the networks and organizations they created to support them, they employed ‘policies and strategies … robust to many alternative futures’. ⁴⁹ The better the hazards of Atlantic shipping were known – and mitigated – the more of those hazards could be, at least in part, factored out of ship design, construction and outfitting, but the fact that late-eighteenth-century Atlantic merchant ships, while incorporating elements of Dutch economy in their hulls, and powered by rigs requiring less labour than alternatives, were still built and armed more heavily than Dutch fluits in the Baltic reminds us that those building and sailing these vessels acknowledged the reality of the radical uncertainties, natural and human, still thrown at them. At the same time, shipping interests were clearly adapting their vessels in the pursuit of profit, both on individual voyages and over time. Success in this required technological conservatism.

While the primary goal of the study of history should be the advancement of empathy, we do study it to inform our decisions about the present and the future – and we should, provided we take pains to tease out, as best we can, the differences between the conditions that produced historical events and the conditions in which we find ourselves. Taking such pains has an important extra benefit: in and of itself, it guarantees us a better-developed understanding of our own situation. When it comes to technology, we are obsessed with innovation, and accustomed to a hyperactive pace of innovation, thanks to the pace of development of the digital computer and its integration into so many pre-existing technologies, and we tend to take a reflexively dim view of technological conservatism. It is worth considering the possible connection between that prejudice and the argument of Kay and King that we have erroneously banished radical uncertainty from its rightful place in our apprehension of reality – that we have assumed we can know the future to an extent that we certainly cannot. Our models, whether of economic processes or the weather or the physics of the ocean in motion (and thus of the behaviour of proposed hull designs in it), are necessarily imperfect. We do not understand all the factors influencing the behaviour of these processes and, to some extent, we never can. We should not, therefore, discount the precious value of trial and experience, which was the only way early modern shipwrights and mariners could learn to predict what would happen on a ship at sea. Technological conservatism prioritizes confidence in trial and experience over confidence in predictive models. As useful as those models have been, since the mid nineteenth century, it is still true, as Kay and King put it, referring to the development of modern jet aircraft, that ‘the only way to resolve the mystery of what happens to a metal tube flying at 500 mph [miles per hour] at 35,000 feet is to try it’. ⁵⁰ We have learned much and we can predict more than they could, but ‘all the improvements in aircraft are written in blood’. ⁵¹ Just as Kay and King urge economists and financiers to recognize radical uncertainty for what it is, and to employ ‘policies and strategies … robust to many alternative futures’, perhaps we should pay more attention to technological conservatism, to incremental, careful adaptation, to the potential for unforeseen consequences of our decisions and to how we can keep our technology best suited to changing needs without setting ourselves up for blindside disasters. The success of the early modern British Atlantic merchant ship, underappreciated as it is, is a useful case study in how that can work.