Determining how quickly a character may learn a skill from scratch in terms of the game is a difficult thing. I am of the opinion that regarding the character's class, the groundwork has already been laid. The ranger who "suddenly" learns druid spells or mage spells was in fact taught both those things during his or her initial training as a lad. Gaining them at a certain level isn't based on the ranger suddenly "knowing" ... it specifies that moment of clarity, when the scales fall from the ranger's eyes and it is at last understood what the instructor was trying to beat into the pupil's head. There are many professions that experience this kind of leap. For me, that's what a "level" is ... an epiphany.
On the other hand, where it comes to a skill that the player knows absolutely nothing about, that is another matter entirely.
For example, a farmer who wishes to become a sailor; a fighter who wants to learn how to read; an assassin who decides to become a deep miner.
These are not skills that are simply learned from experience. They have to be taught, as well ... or else the wannabe miner is just going to dig worthless rock for the rest of his or her life. Worst of the worst, these are also not skills that can be learned while adventuring and leaping about the countryside - they take TIME. A lot of time.
It may be possible to learn to read haltingly, piecing the words together from text, within a few weeks. The vocabulary won't be there, and it's going to be like pain to figure out all the various elements of speech ... but one could piece through an easy text and be understood. Writing is quite a bit harder, but one could scrawl purposefully within the second month.
It is a LONG, LONG leap from there to being able to infuse magic into the written word of scrollmaking, with complex multi-cultural terms and the need for absolutely precise application of symbols and spelling. Scroll-manufacture would require the highest level of ability - and that is not gained in a couple of months. Thankfully, most every spellcaster knows how to read & write ... except that it's possible not to know it, and still be a caster, and my world has that possibility included in the character background.
Several years is more like it. I make the assumption that all classes begin to receive their first training by age 11 ... which means the easiest and simplest class, fighter, still takes a minimum of four years training to become adept at first level. A cleric requires 8 years; a mage, 13 years. And the averages are greater.
Most skills and trades take years to learn. One does not make good barrels or even good beer without a lot of practice.
You may be pressed aboard ship without any previous knowledge, but you're going to spend your first year aboard a schooner learning to clean, polish, serve food, pull ropes for which you don't know the purpose and generally keeping the hell out of the way - if you don't get yourself killed doing something stupid. Most who were pressed aboard ships were found in parts of the city where it was likely to find men with previous experience - and whose hands could be looked at to see how soft they were. Hollywood loved the idea of a rich man being forced to work with the scum ... but chances are such a man would have been turned loose. A month at sea would have left his soft hands split open with cuts and blisters, making him a worthless eater of food.
The problem in a game, of course, is that the players won't be interested in any skill it takes years to learn. In 30+ years of play, I have twice had players who were willing to just sit down for six months or more and do nothing. There seems to be a strange resistance against it. Why shouldn't the party decide to spend six months becoming at least partially adept at sailing or farming. It would make sense to me - raise some capital, buy a sixty acre farm at the beginning of a season and work it until fall. Do it in the space of one night's running - producing a ton of food, making contacts in the area, storing the extra for future campaigns, etc. No one ever does it.
Of course a player could learn to be a sailor; if he or she were willing to take the time. The way I see it, any system would have to be worked upon a bell curve - at the beginning, very little is known, but it is easy to learn things. Near the end, only a few precise things need to be known ... but before you know everything, you DO know most everything. You're not perfect, but you're mostly perfect.
Writing through this post, I think I've just hit on the perfect solution. I had an idea before, but this is better.
Let us say that any particular skill requires the learning of 216 things. It is an ad hoc number, but it's suitably high and that's good - and I have a deliberate reason to choose that number. Let's further suppose that a perfect student would learn all 216 things within the space of 216 weeks. That's a total of 4 years and 8 weeks. Finally, let's suggest that for some things, a week need not be the measure. For some things, two days might be the measure.
We're going to consider a caster learning to read, so were going to consider half of 216 weeks as a good healthy time frame. For this, we're assuming the character hires a tutor and spends between 6-10 hours a week training. A similar amount of time is spent "practicing" ... so the time cannot simply be doubled and everything learned in half the time. The brain cannot be pushed past its usual limit, and the above numbers suggest the time frame a university uses to train a person. Remember, we're not talking about learning to read at a grade one level - we're talking about someone becoming erudite and able to speak as a university graduate.
Consider 3d6, which produces 216 possible combinations. The chance of rolling a 3, in which all three dice show one pip, is 1 in 216. The chance of rolling a 4, on the other hand, where one of the three dice can show two pips, is 3 in 216. Here's the complete odds for each result shown below.
Suppose that for each week that passes, the player rolls 3d6 twice ... and then marks the appropriate slots in the above table (the bottom line would remain blank until filled). Every time the die was rolled in a slot that had not reached its maximum, the character would "learn" ... with the total number of successful weeks added together and divided by 216
in order to produce a %.
However, if the character has already rolled the number 12 a total of 25 times, that person that week fails to learn anything.
The nice thing about this system is that a character could easily become fairly adept at the practice of reading in the space of a year. However, as the spaces fill up, it gets harder and harder to fill those gaps.
Note also that intelligence and wisdom are not considered AT ALL in determining success. For this, I point out that in the game, a fighter with an 8 wisdom and a fighter with an 18 wisdom both fight upon the exact same combat table. They have the same chance at having hit points, and they cause the same damage. In the game, intelligence and wisdom do not determine the values of skills! In the game, intelligence and wisdom ARE skills - they do not measure or affect any other skills in the game. Therefore, they should not affect learning to read or any other gained ability, as they do not affect any previously existing gained ability IN THE GAME.
So please, those of you reading this in any of my three campaigns, do not ask how your intelligence makes it easier for you to be a sailor.