Interesting thread. In my book DESIGNING AND BUILDING SPECIAL CARS (I'm not linking it as it was sold out the day before it was published so the only copies available are second-hand and oily from pawing by hotrodders, and anyway it betrays an earlier fascination with automobiles I have grown out of) I show characteristic graphics of tyre response to various inputs from car, driver and road. The significant thing about these graphs is that they are not a straight line, a convex or concave line, or even a normal distribution evidenced as a bell curve. Instead they are invariably, whatever you can think of measuring, the curve of a complex algorithm, seen visually at a glance as an S-shaped response line lying at various angles to the axes and across the axes. In short, tyre response is the
least predictable of all the vehicle reactions.
A bicycle is a simpler mechanism, but dynamically the tyre and its response is still by far the most complicated thing about it. Fortunately most bicycles don't ever go fast enough for the rider to be faced with the ultimate limit of his tyres. At
In praise of riding low pressure tyres fast, I describe the behavior of Schwalbe's Big Apples in extreme road conditions, and at the edge of handling (
handling is how well the bike recovers when the rider is foolish or unlucky enough to go beyond the natural
roadholding limits of the frame/tyre combination/load distribution).
Before you even get to dynamics, tyres have compounds that are deliberately engineered to respond in a different manner. My fave Big Apples at any one time has three or so different compounds, three or so different anti-puncture bands, several different sidewall constructions, and clearly, by this sample of what is known, probably as many unknown internal differences. All these matters are inputs to a major uncertainty about how tyres of even the same name and size and external appearance will respond to rider and road stimuli. Jan Heine's attempt to overcome the impossibility of full understanding by a "15% rim-drop at the right inflation" rule-of-thumb has been mentioned by Dan but is a bit inconvenient to apply once your wife takes her hemming measure back. It is more of a baseline than a final answer. You'd still need to discover empirically (that is, by riding the bike under a variety of conditions) how anti-puncture bands of different materials react with different inflation regimes.
I've explained the uncertainties inherent in tyres at some length, because it is to avoid aggravating them that I always fit tyres that are as close to perfectly matched as I can get both front and back; I achieve this by buying replacements and spares all at the same time.
Fitting the fatter tyre at the back of the bike is likely to make it livelier but also more dangerous at speed, as it will tend to overdrive the bike and thus make it oversteer. Jobst Brand, the Porsche engineer behind the Avocets which were the first everyday slicks to be sold to Americans, used to say that once you've lost control of the front wheel you're a passenger in your accident; nothing more you can do except pray. However, I have one bike with very lazy geometry that naturally understeers at speed, which is a safety factor at 70kph on badly surfaced, twisty downhills often with loose gravel swept in from the farmers' side roads; it was designed to do that, and everything I have done to the bike since it was new was intended to maintain this comfortable safety margin regardless of what we might meet. But let's say I want to turn this safe, fast tourer into a comfortable commuter. So, on this bike I would fit the fatter tyre at the back and nip in and out of the traffic on the thinner front tyre. You have to know how the bike holds the road and handles excessive inputs before you can start tailoring it. On a bike which already oversteers (bikes with excessively short wheelbases and unnaturally steep geometry: I have no such boy racer appendages), I would put the fat tyre on the front to tame the bike's road manners a little by slowing down the response.
That applies to a thin and a fat tyre of the same make and general construction from the same materials. Adding further complication besides width makes the whole thing too multi-varianced to conclude anything except that you have discovered a change but that you don't know in which direction to go (see the S-bends above) to enhance or reduce the effect. It could be expensive to buy enough tyres to control changes to one variable at a time, so it's probably smart to stay within one manufacturer's catalogue.
I caught "ceteris paribus" above or in one of the links. It's Latin for, "Remember to change only one thing at a time, keeping all others the same."
***
I've often wondered if there isn't a default case for making the rim and tyre on touring bikes wider at the back where the heaviest weight is carried. Not on the short wheelbase daytrippers but on the heftier long-range tourers. It would probably mean supplying two rear wheels with the new bike, one for touring and one for commuting and short rides, which would make most manufacturers cringe.
