Magic in G&G will be divided into four lists: Mind, Matter, Divine, and Nature.
Magic of mind includes all spells that enhance, manipulate, or control the mind, including illusions, charm, telepathy, telekinesis, and sleep spells.
Magic of matter includes spells that create, destroy, or alter living and non-living matter, including polymorph, transmutation, size changing, and elemental magic.
Divine magic encompasses effects considered holy or unholy, such as detecting and protecting from evil or good, blessings and curses, and spells that directly affect life and death, such as healing spells.
Nature magic includes spells that affect plants, animals, and weather, plus healing magic.
Obviously there is some overlap between lists.
Mage characters normally have access to Mind and Matter, though they may opt to give up one of those lists in exchange for learning Divine or Nature spells. Clerics have access to either Divine or Nature magic, but not both.
Spell lists use the terminology "order" or "order of magnitude" rather than "level" for spells of differing power, to avoid confusion with other uses of the word "level." Lists run from cantrips and orisons (Zero Order, usuable at will) to 6th Order.
You won't find many direct damage spells in these lists, though some will have that potential. You won't find, say, the old standby fireball spell. Instead, there are some spells to manipulate fire, and maybe one that will allow the caster to make an existing fire explode outward, extinguishing itself in the process. Damage and area of effect will depend on the size of the fire rather than the caster's level. Getting the target creatures near the fire, or the fire near them, is left to the player's ingenuity.
Spell-casting is similar to the B/X standard, but a character can memorize one spell per day per level of experience, modified by Intelligence (mage) or Wisdom (cleric) adjustments. In the case of penalties, a minimum of one spell may always be memorized. This spell allotment may be divided among the orders of spells usable however desired. For instance, a fifth-level mage with an Intelligence adjustment of +1 may memorize any six spells, dividing them between 1st, 2nd, and 3rd order as the player wishes. The spell table determines the number of spells of each order which the character may cast per day, but any memorized spell may be cast as often as desired within those limits.
Whenever a caster gains access to spells of a new order of magnitude, the first such spell is free. Additional spells may be learned through research.
Learning additional spells is done between adventures, and requires time and money. The base cost is 1,000 sp and five days of time per order of magnitude of the spell. For spells of Mind and Matter the cost is spent on books, reagents, and laboratory time. For Divine spells, the cost is spent on prayer and activities beneficial to the deity or church, such as donations of money, texts, holy items, and so on. For Nature spells, the cost may be expended on books, rare materials, and time spent observing and communing with the natural world.
Spell research succeeds on a 1d20 roll of 10 or greater. Add the character's Intelligence (Mind or Matter spells) or Wisdom (Divine or Nature) to the roll, and subtract 2 for each order of magnitude of the new spell. A natural roll of 1 always fails, and a natural 20 always succeeds. An adjusted roll of 0 or less means that the character cannot attempt to research that spell again without assistance (see below); otherwise, each unsuccessful effort adds a cumulative +1 adjustment to each further attempt to learn the spell.
Learning a spell is easier if it is similar to one the character already knows. "Similar" is left to the discretion of the player and GM. Use the highest order of similar spell known: Subtract 500 sp and 2 days from the cost for each order of magnitude of the known spell, to a minimum of 500 sp and 1 day. Add +1 to the research roll for each order of magnitude of the known spell.
Having a scroll or spell book containing the desired spell increases chances of success by +2, regardless of the order of magnitude.
Assistants improve the chances of success. An assistant must be able to cast spells from the same list, but need not be able to cast spells of the order being researched. A full-time assistant who is not capable of casting the order of spell being researched adds +1. Consulting once per day with someone capable of casting at least that order of magic also adds +1, while studying full-time with such a caster adds +3. Apprentices and masters will generally perform this function for no additional charge. Otherwise, the other caster must be paid 100 sp per level per day for full-time service and 10 sp per level per day for consultation.
So, if a mage with Intelligence 15 wants to research a 3rd order spell, and knows a 2nd order spell that the GM agrees is similar, the cost is:
3,000 sp and 15 days for a 3rd order spell, -1,000 sp and 4 days for knowing the 2nd order spell already = 2,000 sp and 11 days.
The 1d20 roll is modified by -6 for the desired 3rd order spell, +2 for the known 2nd order spell, +1 for Int bonus = -3. The research will thus succeed on a roll of 13 or better. On a roll of 3 or less, the caster won't be able to research the spell again without the aid of another caster.
If the caster had captured a spell book with the desired spell, and was working with an apprentice, she would gain an additional +3 to the roll.
Of course, in a sense time is money, and to an extent the two are
interchangeable in spell research. One may be reduced by a factor of 2,
3, or 4 by increasing the other a corresponding amount. (Always apply these adjustments after adjustments for knowing similar spells.) A character
researching a spell with a base cost of 1,000 sp and 5 days may take 20
days and pay only 250 sp, or may speed the process up, paying 4,000 sp
and finishing in 1 1/4 days, for instance.
The aim of these rules is to encourage a degree of specialization without a load of rules for specialist mages. Of course, a character may choose a scattershot approach to learning new spells, but the reduced cost and improved odds of success for learning spells similar to what one already knows provide an incentive to pursue the path of least resistance.