Friday, 31 August 2007

A review of Stout by Michael J. Lewis

I've already told you that I have a huge collection of books on beer-related topics. Some of them I've even read. I tend to avoid modern general books about beer, because I either already know what they say or disagree with it violently. Brewery histories and brewing manuals are much better for my eductaion and my blood pressure.

This week I ordered another dozen books: "Tijdschrift voor Brouwerij en Mouterij 1939-1940", "Münchens bryggeri 1855-1910", "The Manual of British and Foreign Brewery Companies, 1947", "Manbre & Garton Limited 1855-1955: A Hundred Years of Progress", to name just a few. I don't know why I'm telling you this. That list's putting even me to sleep.

Why have I written a review of "Stout"? A couple of reasons. Firstly, because there are so many ludicrous assertions in it. Secondly, because I've seen it quoted as a source several times. It scares me that anyone could take this book seriously, much less use as a starting point for further study. I hope the following will explain why.

Stout by Michael J. Lewis
The Classic Beer Style Series is a noble attempt to document the world's classic types. Its volumes have covered the most widespread British, German and Belgian styles. Designed with the homebrewer in mind, they discuss recipes and brewing techniques along with the history and characteristics ogf each beer.

I own most of the series and there isn't one from which I haven't picked up some useful snippets of information. Even Michael Lewis's "Stout", which at times had me frothing at the mouth.

The book is divided into six chapters: 1. Introduction, 2. The Origin of Stouts, 3. Commercial Brewing of Stouts, 4. A Taste of Stout, 5. Our Survey of Stout Brewers, 6. Brewing Stouts at Home. It would love to report that they were all of equal merit, but that is sadly not the case.

The good bits
Let's start with the positive. As you might expect from an academic with decades of experience in teaching brewing science, the author is very strong on the technical aspects. The chapter on commercial brewing has some interesting detail about London and Dublin water. The description of the brewing process, though effectively limited to a description of Guinness in Dublin, also provides some useful insights.

The most interesting original material comes from the survey responses of breweries from around the world in Chapter 5. The survey covers the following topics: raw materials, mashing practices, kettle boiling, fermentation and conditioning, package and dispense, physical and chemical analysis. Nowhere else have I ever seen such detailed information about commercial brewing practices from. The breweries who responded are a very diverse bunch both in size and geographical location: SAB (South Africa), Banks (Barbados), Carlton (Australia), Castlemaine (Australia), Desnoe abd Geddes (Jamaica), Kirin (Japan), Beamish & Crawford (Ireland), Asia-Pacific Breweries (Singapore), Hales Ales, Marin Brewing and Pike Place (all USA). For me, the book is worth buying for this chapter alone.

As an unashamed numbers junkie, the tables of analyses of various stouts (detailing OG, ABV, colour, calories) are also fascinating. They cover both modern and historical beers. It's difficult to complain about the quality and quantity of raw information packed into the 171 pages.

The less good bits
What had me chewing the carpet were the discussions of the history, characteristics and stylistic variations of stout. The opinions expressed on these topics range from the bizarre to the ridiculous.

Chapter 2. The Origin of Stouts
Lewis is convinced that stouts have always been black:

the idea that stouts were not black beers, as today, again begs the question,
Who says so? The implication is that such beers were somehow impossible. But
this cannot be. . . . Coffee roasting and grain roasting are exactly parallel
technologies and it wouldn't surprise me if some clever brewer hadn't thought to
compete with coffee by using roasted grains
.” (p.7)

This is a wonderful piece of revisionism, arbitrarily moving the invention of patent malt back a century, without any hard evidence and purely for the convenience of the author. Its main purpose is to rewrite history in such a way so as to allow Stout to be a black beer, predating porter and quite distinct from it. Let's consider the reasons why it's total crap:

  1. That the first stouts weren't black is sort of given away by their original name: Brown Stout.
  2. Until around 1800, when brewers started to use pale malt in Porter, it was a 100% brown malt brew and there was no need for anything extra to colour it.
  3. The roasting of grains - patent malt - was, as the name hints, a new idea when it was invented in 1817. That's why a patent was granted. There were very specific circumstances - the prohibition in 1816 of any colouring agent in beer other than malt - which led to the invention of patent malt. Brewers desperately needed an alternative colouring material.
Though he acknowledges Stout's origin in London, Lewis seems to be under the illusion that most English breweries copied the style from Ireland. He asserts on page 22:

Although I earlier argued the case for stout preceding porter, there is clear
evidence that commercial brewing of stout in England grew out of the wide
popularity of porter in London and elsewhere. The ever increasing importation of
Guinness to the English market during the mid- to late nineteenth century surely
drove many English brewers to emulate the Irish style. For the most part,
however, the stouts made in England and Scotland during this time were not dry
like Irish stout but sweet, many containing portions of milk sugar

Let me get this straight, the popularity of Guinness prompted English brewers to make a stout, but they made it in a totally different style? “People seem to like that Guinness. I know - I'll brew a Stout totally unlike it. It''s bound to sell sell like hot cakes.” they must have thought. The reality was somewhat different. Throughout the 1800's most English brewers produced a range of three or four Porters and Stouts. Were these all imitating Guinness?

Here's a typical example of a brewery's range of draught beers:

Bradley & Co, Soho Brewery Sheffield 1870
Mild X Ale
Mild XX Ale
Mild XXX Ale
Mild XXXX Ale
No.3 Australian Strong Ale
Pale India Ale
Stout Porter
XX Stout Porter
Strong Brown Stout

Guinness, like Bass, was a premium bottled product, sold in pubs tied to other breweries. This situation existed up until at least WW I, as this article demonstrates. Sweet stout, a low-gravity beer laced with unfermentable lactose, didn't appear until Edwardian times. Even then, English brewers continued to make London-style draught stouts, which were part of the standard range in pubs. London breweries were brewing one, and sometimes two, draught Stouts until at least WW II. In 1936 Barclay Perkins brewed a Porter and seven different Stouts, at strengths varying from 4.4% to 9% ABV. If he had bothered to do even the most rudimentary research he wouldn't come up with such total crap. Strangely enough, I would expect someone writing a book on the subject to have consulted some original documents. The brewing logs of three of the great London Porter/Stout breweries - Whitbread, Barclay Perkins, Truman, Hanbury and Buxton - survive for the period 1805 - 1970. More than enough hard facts for anyone who can be arsed to look for them. But why bother when guesswork and third-hand stories will do?

Lewis seems unaware that, in the period he refers to, Guinness Extra Stout was completely different to the beer that bears the same name today. In 1896 it had a gravity of over 1070º, now it's around 1042º. Despite the fact that these figures appear in the book (pages 35 and 121). "Irish Dry Stout", as style nazis like to call it, didn't appear until after WW I. Before 1917 it would have tasted much like Guinness Foreign Extra Stout - a totally different beast.

The plain crazy

Chapter 4. A Taste of Stout
Here it becomes clear why Lewis was determined to make early London Stouts, in defiance of the known facts, black. This is how he categorises a Stout on page 66:

to conclude, it was not difficult for us to decide that a stout is simply a
black beer called a stout by the brewer who made it
Now there's a useful definition! I'll expand later on precisely why it's of no use to anyone. I first want to observe that the only distinctive feature of a Stout, in his opinion, is its black colour. No wonder he was so keen on making the early London beers black - if they hadn't been, then they wouldn't have been Stouts, would they?

Why does he propose such a vague formula? Ostensibly, it is the result of a systematic tasting of stouts by a panel. Analysis of the flavour profiles produced by the tastings failed to show any discernible patterns which would identify subtypes. So, that's it proved then. There's only one type of Stout. How stupid I've been all these years thinking that Draught Guinness and Courage Russian Stout were different subtypes!

But hang on a moment, this doesn't agree with my own, admittedly subjective, observations. I am able to identify patterns in the characteristics of Stouts and even - god forbid - substyles. Why was Lewis unable to find them? Some comments earlier in the book hint at the reasons. P.33:

I'm a lumper and my tasting experience tells me some stouts are sweet tasting
(including some milk, cream and oatmeal stouts) and others are not (dry stouts).
Further classification does not help me much
A cynic will be tempted to think that he didn't identify any substyles because he started his experiment convinced of that fact. Subconsciously constructing experiments so that they produce the result you want or expect is a known phenomenon. A closer examination of Lewis's tastings reveal why they were never going to provide evidence of subtypes.

23 beers were tasted, broken down by these regions of origin: USA 10, Australia 4, Jamaica 1, UK 7, Japan 1. The following subtypes were claimed by the names of the beers: Stout 11, Oatmeal 3, Cream 2, Imperial 2, Porter 2, Extra 1, Triple 1, Black Beer 1.

Does a huge flaw start to become apparent? The sample size of 23 is way too small to draw any meaningful conclusions. With no more than 3 examples of any single subtype, it's farcical to claim to have proved their non-existence.

Yet the ridiculously small number of beers analysed isn't my biggest problem with the experiment. The concept behind it - to find universally applicable definitions - is even more ludicrous. What has been the result? A statement classifying Stout that could scarcely be more vague, yet which still manages to exclude the original beers in the style! It demonstrates perfectly the intrinsic futility of the endeavour.

There is one set of figures Lewis employs which do show a definte pattern grouping beers together. It doesn't come from his own tastings, but from information provided by Guinness. This data - and Lewis's dismissal of it - illustrated to me what was wrong with his whole approach. It's an analysis of the properties of the draught Irish stouts. They aren't identified, but I think that we can safely assume they are Guinness, Murphy's and Beamish. He recognises the similarities, but remarks (p.81):

Since these three stouts compete head to head in the same market for the same
consumers, some similarities among them might well be anticipated
He doesn't realise it, but he's stumbled upon the only meaningful method of classifying beers - based on their context.

Let me explain what I mean by this. There are several factors which influence the nature of beers in a particular place at a particular time - taxation, legislation, local tastes, raw materials, climate.These factors are not constant over time and location. Thus a style, for example Stout, cannot be expected to be the same everywhere in the worlld.

Here's a good example. Carnegie Porter was first brewed in Sweden as an imitation of a British Imperial Stout. It must have originally been at least 8% alcohol. In the 20th century the strict and sometimes bizarre alcohol legislation in Sweden forced it down to 5.6%. Rather than just drop the beer, the brewer made a spirited attempt to retain as much Imperial Stout character as was possible. Now if you compare it to Courage Russian Stout, there are some pretty obvious major differences, yet it is not wholly dissimilar.

Using Lewis's approach, this would demonstrate the meaningless of the category Imperial Stout. I see it in a completely different way: it highlights the inadequacy of seeking single a definition to cover beers existing in very different legal and social environments. lf we admit that the characteristics of a style will change over time and that this change will be driven by local circumstances, then the heterogenous nature of Stouts today should not come as any surprise. Porter and Stout were the first international styles and have been brewed outside the UK (for example, in the USA) for more than 200 years. Is it reasonable to expect that a modern American Stout would display the same features as either an 18th century London Stout or a modern Irish Stout?. Of course not. No more than we would anticipate the accent of a Londoner to be the same as that of a New Yorker or a Dubliner. Yet it doesn't mean that they aren't all speaking English.

Taking “context” into account - limiting definitions to a particular time and place - then it is possible to identify subtypes, as Lewis himself was able to do with draught Irish Stouts. Taking a random sample of widely differing types of Stout from around the world and trying to find clear patterns is so obviously stupid, that you wonder why anyone would want do it. If he hadn't rewritten history, Lewis would have been forced to coin an even looser definition, namely that “Stout is a black or brown beer called a Stout”.

In a way, I'm surprised that there are any beer styles at all in Lewis's world. His methodology, if strictly applied, would lead to just pale, amber, brown and black beer. Universally applicable but - for both consumers and brewers - totally useless classifications.

The length and detail with which he describes his extremely unscientific and pointless experiment (23 pages compared to just 32 on the whole history of Stout) are a clear indication of how important the topic of non-style is to him. You might expect a book dedicated to Stout to describe its many variations, rather than taking so much trouble to prove that they don't exist.

Despite my reservations about some chapters, this is still a useful book. It's great strength is the sheer volume of information in contains. This compensates for the dodgy analysis of the author. But the historical sections are pure fantasy with no basis in fact. It would be funny if it weren't for the fact that people use this book as a serious reference work. Try not to be too distractred by some of his more dubious arguments and use the data to draw your own conclusions.

Stout (Classic Beer Style Series)
Michael J. Lewis
Brewers Publications,U.S
ISBN 0-937381-44-6


    GenX at 40 said...

    See, this is why I encouraged you to blog. This surpasses my patience and my focus. Perhaps your work are putting the "mmm!" in manic but that is a complement of the highest order.

    Tell me this, though. Would a good bit of industrial archeology as well as primary record research put questions like this entirely to bed? For example, with an entirely different style, I have my questions about claims to 3 year or more pre-Victorian era storage for non-monastic Belgian sour beers but showing me the spots and massive cellers which must have been used would smooth my furrowed brow. Is such archeological reportage available to a mere mortal like me?

    A Good Beer Blog

    Ron Pattinson said...

    Alan, glad to hear you've liked my blog so far. If only I weren't spending 3 hours a day writing it.

    Good question you asked. I know that there have been archeological investigation of brewing sites in the Low Countries. I seem to remember Unger's huge volume on the history of Dutch brewing includes some.

    I've never seen a similar work for Belgian brewing. Jef van Steen's books seem well researched and there is one on lambic beers (I have a copy of it in my book pile). But I'm not sure if they have been translated into English or not.

    To be honest, I wouldn't know where to look for archeological publications. One word of warning though. Podge's girlfriend is an Egyptologist and she told me about a site she had visited in Egypt that had been identified as an ancient brewery. She was very dubious because the stone troughs looked too small and shallow to have been much use for brewing. This is the problem: archeologists don't necessarily understand much about brewing

    Which particular Belgian styles were you interested in: Oud Bruin and Lambic? You're right to say that there must have be some physical evidence if beer was stored for a period of years.

    GenX at 40 said...

    Note the post today at Zythophile on the use of the word "stout" in 1700s brewing:

    Ron Pattinson said...

    Thanks for pointing that out. It's an excellent post about Stout. He really knows his stuff.