I like the way you describe the cross path synergy system.  It sounds like it would add incentives for people to research all fields, rather than a specific one.  However, I don't see how the synergies will create more strategic decisions when researching.  Firstly because to my knowledge, the tech tree is semi random which means that if you get lucky you might get to play with crossbows (your example), but this is purely luck, not did I decide to research that.  Secondly, if the tech tree isn't too randomly populated, using synergies like the crossbow will still cause the issue mentioned by lambdaman, "you end up with basically one research tree that everyone follows".

While I'm on the topic of synergies, I think that synergies shouldn't give more "powerful" abilities/weapons/armours/spells, rather give abilities/weapons/.... that are the same"power" but act differently.  Basically give you an additional strategic option on how to behave (unit design etc).

I think the strategic options will come about more in the battle beta (4 I think).  As in most TBS games I've played in the past, the tech tree really revolves around two areas: Fighting and city building.  Fighting is where most strategic options are created, using your example above about the shield wall concept, can only come into play with tactical battles.  Therefore I believe the key to having strategic options in researching is having enough combat modifying technologies.  By this I don't mean, my pointy stick is bigger than your pointy stick and does more damage.  I mean this unit has X equipment which makes it act radically different from your man with the bigger pointy stick.  For example, see MoO2.  There are technologies like the time warp facilitator (act twice per turn at the cost of more armaments/defense), Inertial Nullifier (higher defense, further movement, at the cost of more armaments/other specials).  Black Hole Generator (holds target immobile for 5 turns then auto implode, at the cost of direct damage weapons or defense), damper field (damage reduction at the expense of hit points ).   I beleive these kind of choices are what you are driving at.  and I agree!

To sum up, I believe there needs to be certain mix of technologies:

1) The stock standard, my stick is bigger now technologies.

2) The synergy system as described above (not more powerful, but new strategic option).

3) Large amount of modifying techs.  These are the "fun" ones which give you the strategic options.

4) Everything has to have marginal cost.  ie, there has t be a limit on what can go in a unit design, make you CHOOSE what to equip, not just make uber units.

Given, a lot of modifying techs will be in te magic tree, but that shouldn't stop there being plenty of modifying techs that are mundane.  (choosing full plate armour over leather armour SHOULD be a strategic choice, however in the current battle system this is not the case as it is incomplete).

PS: Uber units are the bane of strategic choices as it resorts to who has the biggest stick/teched up the fastest....

I like the idea of cross path research, a great many things are possible with that.

I think that a big factor in research will be the actual usefulness of the various technologies available.  In a later beta, having magic, diplomacy and more incorporated into the game rather than being placeholders, we should get a better feel for how useful each of the techs becomes.  I would rather have fewer more meaningful choices.  As long as the majority of techs are useful there should be room for many variations in gameplay. 

I would like to see an early choice that guides the thinking/focus of a particular culture along a consistent "theme.".  This would have 2 main effects.  First, a culture that is more militarily oriented to using horses versus a culture that is more infantry oriented will naturally take a research path that helps their orientation.  For example, the riders may look at horsebreeding and develop different types of horses for scouting, heavy cav, light cav, etc.  Meanwhile the infantry culture might look at different styles  such as swords (even the different types of swords), spears/pikes, axes/maces.

Of course, both cultures will still develop/discover similar techs.  However, how these are used would be utilized by the different cultures to advance their specific "theme." For example, an advanced metal is now available.  The horsers use it to make better stirrups, thus improving their movement/agility.  The groundpounders use it to create better body armor, or shields, or weapons. 

I agree with most things you posted, and particularly agree with this quote here, since it captures the essence of what I mean: NOT (only) bigger and better as options, BUT rather different, for example, in terms of combat: Stealth or fear instead of pointy stick; tactically, but not universally, useful shield wall instead of "better" armor; etc.

With synergetic cross-path researching, I mean that this would introduce a method of adding differences into the game mechanics without making each breakthrough the sole measure of a discrete changes; this might be able to increase strategic choices because it would make various options more viable for players -- in the sense that every tech tree, and many breakthrough milestones on each tech tree, become enhanced in value for every player at more stages of the game. In other words: It would help avoid the situations we know from other games (be they RPG or TBS) in which investing in a particular tree becomes meaningless once a certain set of conditions has been met, for example: (RPG) Once your character has purchased X or more "dual-wielding" skills, it becomes increasingly meaningless for that player to purchase any "archery" skills, thus narrowing a player's set of valid strategic choices; (TBS) once your city has built X or more military production buildings, building economic boosting buildings becomes increasingly meaningless. With more synergies, we can expect a player to have invested heavily in Warfare to be rewarded by having big strong armies, and his options will also still be varied, in that she might choose to invest in few squads of elite horse soldiers or large squads of decent archers; yet the players will have more "flavors" to choose from if, in addition, their research into Civilization or Diplomacy etc also add different flavors to their warfare path they can, but need not choose from, such as some I suggested (battle priests, spies, etc.) -- not just pointier sticks. This would also reduce the chances that all players continuously beeline for one particular path to master because this is the "best" option.

Additionally, this would also increase the net amount of variations possible in-game, and thus make late-game combinations less foreseeable; for example, some particularly elite abilities I suggested earlier

e.g.1: breakthrough Advanced Charming II (Magic path): lets you cast a hard-to-resist combat charm spell called Compulsion on most Kingdom humans, a high-level breakthrough, as well as a non-combat charm spell Conviction used only on envoys in friendly cities; but if you ALSO have  Subtle Talking II (a mid- to high-level Diplomacy breakthrough, which itself gives you the ability to negotiate MPPs with neighbors at a high bonus) then the breakthrough Advanced Charming II might ALSO gain an additional effect, e.g. a substantial bonus to the spell Compulsion, and the ability to cast Conviction with a bonus as well as in all cities, not just friendly ones; but IF you ALSO have the breakthrough Advanced Animal Speech (high-level Adventure path), then you may cast Compulsion on ALL entities of all factions, including animals.

Just an example. What I am saying is that (1) during the later game, people might still be getting interesting benefits from discovering even lower-level techs from unspecialized trees often considered to be "wasted" in other late-game scenarios (thus also helping to avoid late game tedium), but also (2) later game permutations will be increasingly interesting, varied, and unlike the my-stick-is-bigger-than-your-stick foreseeabilitiy; and (3) the strategy of technology-pushing (that is: concentrating lots of resources into research) into various branches becomes viable for rare multi-path possibilities that not everybody will be able to obtain (e.g. abilities which require at least 3 different mid+ or high-level breakthroughs) in the course of one single game incarnation.  

Moreover, it might also make for less clutter. If you want to have, say, 1,000 "abilities" in the game (actions, statistical boni, empire-wide effects, spells, building options, combat maneuvers, etc.), you might be well off having 1,000 discrete breakthroughs required to enable these, by having about 200 breakthroughs per path. But what if you want to have (or mod) 10,000 different "abilities". Attributing the ability to perform or unlock or build or activate some of these abilities not to one discrete breakthrough, but to more simultaneously, will give you (especially you as a modder) more room. That way, one early breakthrough, say Simple Mining, might give you up to 100 boni, depending on what other breakthroughs you have unlocked and what other conditions apply later in the game.

I would like a tech tree where the low level techs are easy to get, but with costs of researching out of your focus rising exponentially the higher in the tech tree you go.  So while you might choose to have a Roman like focus on architecture (opening up really sweet possibilities for high end buildings), you will only ever reach a medium level in something like metal-working or seafaring.  While I have a Celt/German focus on metalworking, but my people lack the sophisticated architectural skills necessary to build large structures, or the ability to have anything other than mediocre cavalry.  I think this system is good because it gives the player options about where they want to go tech-wise while not opening up the whole tree, and causes significantly different choices to shape your society in a very real, meaningful way.  The more consistently you research a tree and the further along in it you get, the more expensive all the other trees become.

Naturally this assumes that technology trading does not exist, or carries some pretty significant penalties/challenges with it.