Tweet
Social Model v.2
Social Model
I. Introduction
The Clash of Civilizations will be a challenging game not only because of the better AI it will include, but also because the Clash world will be more complex compared to games of this type. Most of models developed in the Clash Project have assumed so far the existence of civ's cultural attributes where they can get inputs from to simulate how different cultures act in different ways and how civilizations can appear to have their own styles and histories based on social effects. It is the main role of this model to produce that necessary social base we're all counting on. This model only describes how people can be characterized through cultural attributes and which mechanisms exist to make attributes change overtime according to the civ's experiences. The effects of culture in the game are supposed to be managed in the rest models. In particular, the Riots Model will take care of major social events such as revolutions, independence attempts, etc.
For the nature of the subject, it's possible for some people to take in the wrong way some opinions about culture and culture evolution. Audiences are encouraged to be open minded when reading the model, because, of course, the only goal here is to replicate what we've seen in history as a general trend and avoid discussions on what's good or bad about cultures.
II. Model Overview
The problems to solve are mainly two: How cultural attributes change as time passes by and how different ethnic groups interact with each other. Modeling the first one is obviously needed, while the importance of the second one needs an explanation: Civs won't interact with each other always like nations, each in its own space. If romans take Egypt, then egyptians end up living under roman rule and they won't necessarily will become roman citizens and support the new regime. Even more, Rome can invade just a part of Egypt, so you can find egyptians in roman provinces and in egyptian civ. In other words, the problem here to solve is the interactions between cultures when more than one occupy the same space (provinces).
The approach I took for the model is this: Each civ can hold several ethnic groups (tribes/nationalities) in it. At a civ level, each of them is described with a set of cultural variables or characteristics. It is assumed that all people with the same nationality have the same characteristics regardless of the place where they live within the civ. This means, FE, that romans in the capital province share the same characteristics that romans living in central Europe provinces. At the province level we only need a dynamic array of the existent nationalities and the populations of each. FE, a roman province in central Europe could be something like this:
Romans-1200
Celts-950000
Franks-1200000
while in a roman province in the greek peninsula would be:
Romans-2600
Greeks-1700000
The number of elements in the arrays change as migrations and other events happen. Populations will be stored on a tile/square basis. Province populations are just the sum of the populations in the tiles forming the province.
The set of ethnic groups (EG) in one civ will be called Population Composition (PC). Each EG in the PC has the following information:
Each turn what the model does is change the characteristics of each ethnic group in the civ according to some rules explained later and where religion plays a fundamental role. The other important factor is govt. It's a govt decision if any ethnic group is above the rest (FE, romans in the roman empire). In general, the govt will determine a level of ethnic discrimination defining how groups will interact with each other. These two factors, religion and govt, and their role in the model are discussed in the next two sub-sections below in general terms.
It should be clear from what has been said above, that "civ" and "tribe" (or ethnic group) are not synonymous. A player plays a civ, not an ethnic group. The player playing the roman empire controls the roman govt and through it, all what happens in the controlled provinces (the civ), but romans are just one of many EGs in the civ. The romans having a special position respect to other tribes and calling the civ "roman civ" is, above all, a govt decision. The civ may indeed have a nationality of its own. FE, a player might be ruling Yugoslavia (civ's nationality=yugoslavian), while EGs in this civ, although generally referred as yugoslavians, have other nationalities (serbs, croats, etc) and there's no EG above the rest. This example also shows that in this model "nationality" means the tribe you feel you belong to, not something imposed or declared by a govt.
II-1. Modeling Religions
Religions are definitely needed in the game. They have given shape to culture for most of (if not all) history and they have also played huge roles in wars and politics, so they're a cool ingredient to add in Clash. The govt model handles how the Church acts in the political arena. In this section it's explained how religion affects EGs.
A religion will be described by several attributes, many of which are the same that describe a culture. The idea is religions "dictate" to their followers how to behave, so they show the people what values cultural attributes should take. FE, if one religion has an aggressiveness of 70, then all the people who follow that religion will tend to have their cultural attribute "aggressiveness" equal to 70. So, the main role of religions in this model is being culture changers, making cultural attributes alike to religion attributes.
There've been many religions and there're still many of them. Their evolution from their primitive forms to how they're today is a complex matter. In order to avoid the game from keeping track of a lot of religions, their birth, evolution and possible disappearance, I've chosen the following approach: The really important religions are what I've called "the Great Religions of the World". There'll be from 3 to 6 of them in the game (the number will be decided by the team, but that should be the order of magnitude). The model will manage how an EG goes from following primitive religions to following one of the GRW. What makes the difference between a primitive religion (PR) and a GRW is the "exportability" of GRW. PRs are extremely associated with the particular ethnic group that practices it, so it's very difficult for others to embrace the same beliefs and therefore these religions cannot spread. Other peoples can sometimes take elements or even the whole cult, but always adapting it to their own nationality (FE, greek religion passing to romans). If you take ancient egyptian religion or aztec beliefs, it's clear to see how religion has to do with the egyptians and aztecs respectively, but not with humanity in general. GRW go beyond the tribe, so no matter where you come from, it's sufficient to follow what the religion dictates to be part of it and therefore it can spread through the world.
GRW will be described with several attributes shown later. The model won't keep track of PRs' attributes, only GRW.
At the beginning of the game EGs will support a PR with 100%. It's assumed that a PR in one EG is different from one in other EG because of the relationship primitive religions have with nationality. PRs have no name (or can be called after the EG, like "egyptian cult" FE).
The game will randomly give birth to each of the GRW after some thousands years of play. Any EG in any civ will be picked as the starting tribe to embrace it. From that point, the new GRW starts to spread as will be explained. Any EG will be allowed to support only one of the GRW. This is not true in real life because in a group of people different cults can coexist, but when we see the world we notice there's always a very dominant religion and some small groups following other beliefs, so I think the restriction I'm using is good enough and it helps keeping the model simple.
For spreading purposes, in each civ there'll be 2*N dummy variables indicating if each of the GRW is known there and if it has followers, where N is the number of GRW in the game. These will be called "Contact Variables" as in "The civ has/hasn't contact with GRW X" and "Acceptance Variables" as in "There are/aren't EGs accepting the GRW X". FE, in one civ there's contact with christianity and Islam, but not with Buddhism, so the contact variables are CHR-1, ISLAM-1, BUD-0. But only Islam is practiced, so Acceptance Variables are CHR-0, ISLAM-1, BUD-0.
Also for spreading purposes, each civ needs to store its level of "social contact" with the rest civs. For simplicity, that level will be given by the number of trade routes between the two civs. That info will be stored in a list (Known Civs List). Each time a trade route is created or deleted, the lists are updated for involved civs.
II-2. The role of government
Usually a civ's govt is related to a particular EG (a nationality) and it's also common for a govt to be related to a religion. The govt model uses the Ethnic Discrimination (ED) Policy for the first type of discrimination. ED can take the following numbers:
0: No discrimination (uncommon. Think Yugoslavia). When ED=0 the civ is said to be "multiethnic" and the player will be asked to introduce a new name for the civ, like Yugoslavia. This is important, not just flavor. As will be shown, EGs with low Nationalism will tend to lose their identity and adopt the civ's nationality, so a player ruling an empire with many EGs with low Nationalism might choose as a strategy to use a multiethnic civ in order to unite the tribes. This is how the model makes possible the rising of new nationalities like the french from gauls, franks, celts and others. However, this strategy wouldn't work in Yugoslavia, where nationalism in the EGs present was high.
1-9: A nationality is preferred (i.e. only one EG) and all others are ethnically discriminated (Minorities). The civ is named after the EG's nationality (e.g. Roman Civ). This is the most common case. Going from 1 to 8 gives different levels for treatment toward minorities, from respect to other tribes and their customs and the possibility to change nationality to the "dominant" one, to very disrespecting treatment (think nazis).
When a govt chooses to be multiethnic and after a new tribe name is picked, the model merges all cultures to create the characteristics for the new people using relative populations as weights in a weighted sum of the original characteristics. This is no sudden forced culture mixing. The new tribe will have zero population. As the game advances, EGs with low Nationalism will slowly drop their current nationality and become part of the new EG.
The govt can also discriminate via religion. If there's a govt with an intolerant official religion, EGs supporting other beliefs can be considered minorities. This is done in the govt model via the Religious Discrimination variable. Since an EG can follow two religions when the EG variable PR% is greater than 0% and lower than 100%, there's ambiguity in knowing if the EG should be discriminated or not. In that case we'll apply the rule of considering the EG as supporting the GRW alone. This is done because as will be shown, once the EG takes a GRW, it'll inevitably embrace it a 100%, so it's just a matter of time to consider the EG as totally committed with that GRW.
Minorities, then, are all EGs discriminated via nationality and/or religion. The govt discriminates and the whole thing about discrimination has as purpose, here in the social model and in the riots model, to change the attitude of each EG and the things affecting them.
III. Cultural Attributes (Characteristics)
These are the cultural attributes and their role in the game. More can be added given the needs in each of the rest models. All variables are in the 0-100 range.
Ethnic Tolerance: It represents how you see other tribes and what kind of relationship you think exists between your tribe and others. A low ET means you feel your tribe is superior and the rest hardly qualify as human beings. A high ET means you respect other tribes and you're willing to live together with them.
Religious Tolerance: This is the tolerance of other religions (by other I mean the ones not followed by the EG). Low scores mean people feel upset about having to live with people with other cults.
Aggressiveness: This defines how aggressive the culture is. A high score indicates a very warlike culture while a low score indicates peaceful. Leaders of an aggressive culture will find it easier to wage war against their neighbors. A very high score could also give a bonus to the efficiency of troops.
Traditionalism: This represents the people's ties to the past. Its role is force culture to stay as it currently is.
Nationalism: Will be defined as "The notion of being a part of something bigger than your local clan (a nation) and the people's willingness to form that nation and live under one same government of their own". This variable was sometimes known as "provincialism/nationalism". Low Nationalism means you care only about your local clan and you don't see much difference between your tribe and others. A high Nationalism shouldn't be confused with a fanatic or blind nationalism. It's only a rather strong feeling of belonging to a specific tribe and a high opposition to be ruled by other tribe. Nationalism is very low at the beginning of the game (thousands of years before Christ), with some random small differences from tribe to tribe (values in the range 1-10).
Land Connection: This is the people's feeling of belonging to the province they occupy. The love for where they live and the idea that the land they occupy belongs to them. Initial LC will be high because the EG is assumed to have lived there for some time. It'll be used in the Riots Model.
Asceticism: This represents the people's desire for material goods. A high score indicates a lack of interest in material goods, while a low score indicates lots are wanted. It has a role in people's behavior in the economic field.
Importance of Religion: The highest this number, the more influenced the rest characteristics are by religion doctrine. This happens because people want to do what religion says. Importance of Religion will be in general very high at the beginning of the game for all cultures and in general terms it will decrease as time passes by.
Individualism: This is the people's value of self. Its main role is modeling what type of economy people wants. The more individualistic, the more inclined to capitalism and vise versa.
Initial values will be given randomly or using pre-set values defined by us and ad hoc to each historical tribe. The player will decide which option to use.
IV. Attributes of GRW
Each GRW has the following cultural attributes: Religious Tolerance, Aggressiveness, Asceticism and Individualism. They represent the moral code the religion tries to impose on its followers. Religion has no say about the other characteristics described above. The following attributes also exist in GRW:
Sacrifice: The higher this score is the more is required from the followers of the religion. This is money and economic resources the clergy takes away from the people that allows the church to act. The economic model should use this variable to know how much people invest in religious infrastructure.
Countdown: Used to determine when the GRW is born. Its initial value in the game is equal to the date the player chooses to start, doubled (FE, 4000BC implies countdown=2*4000=8000).
Holy Land: The tile in which the religion was born (randomly picked at the moment of religion's birth). This is to model crusades and that sort of things.
After the GRW is born, its moral code remains fixed for the game. Religions have shown to be very reticent to change their doctrines, and although we know some changes have occurred, the assumption made here is still very accurate and keeps the model simple. Sacrifice and Holy Land are also fixed and Countdown stays at zero once the religion is born.
Primitive Religions are supposed to be described with the same variables, but the game won't store that info as with GRW. We'll assume that if you take a EG, the PR associated with it has attributes equal to the EG's cultural attributes. In other words, cultural attributes and religion attributes for PRs are indistinguishable. If you want to know what the chinese PR thinks about individualism, just look at the current individualism score of chinese EG. This is made to avoid keeping track of each of the dozens primitive religions existing and their attributes and to give GRW the importance they deserve, being much more autonomous and homogeneous through out the world. As for sacrifice and holy land, which is info we can't read from the EG attributes, holy land is never needed for PRs and Sacrifice will be assumed to be equal to the average between (100-Individualism) and Asceticism.
V. How the model works
Each turn we do the following:
1. Apply the "GRW birth" sub-model to the game.
2. Apply the "Religion Spreading" sub-model to each civ.
3. Apply the "Nationalities" sub-model to each civ.
4. Apply the "Cultural Evolution" sub-model to each civ.
5. Apply the "Migration" sub-model to each civ.
V-1. The "GRW birth" Sub-Model
The countdown variable for each GRW is checked. If it's zero, do nothing. If it's greater than zero, it's decreased by R*L, where R is a random number between 1 and 3 and L is the length of the game turn in years. If after the decrease it's zero or negative, then set it to zero and give birth to the GRW. This means a random tile is picked anywhere in the world. If no population exist in the tile, another is chosen until a tile with people is found. If there's more than one EG in the tile, one is picked randomly. Contact Variable for the GRW in the civ where the EG lives is turned to 1. The religion's moral code is made equal to the corresponding attributes in the EG. Sacrifice is made equal to the average between (100-Individualism) and Asceticism and Holy Land is made equal to the tile's code number.
V-2. The "Religion Spreading" Sub-Model
For each EG in the civ:
First: If the EG has PR%=100% (no GRW is known and accepted), the model looks at the Contact Variables and for each of them equal to "1", the religion is checked for acceptance by the EG. The EG will accept the GRW if the religion's moral code is not too different from EG's culture. A "distance" between culture and religion's moral code is computed as D=square_root( (C1-R1)^2 + ....+ (Cn-Rn)^2 ), where Ck is the EG's value of the kth cultural attribute and Rk is the religious value for the same attribute. If D is less than a given threshold (TBD), it is accepted. The GRW variable in the EG is turned to the name (or code number) of the GRW accepted and the Acceptance Variable for the GRW in the civ is made equal to 1. A GRW spreading rate is then computed for the EG as (1-Traditionalism)*(K1-D)/K2, where K1 and K2 are scale factors allowing to use D in the calculation so the more similar attributes are, the fastest people will take the religion.
If a GRW is at the govt (i.e. it's official), it has more power to convert people. So, instead of comparing D directly with the threshold, we'll really use D*(1-G*RC%), where G is a dummy being "1" if the GRW being checked happens to be the official one and "0" if not. RC% is the political power of the Religious Class, from the govt model.
Second: If PR% is greater than 0%, then PR% is decreased by the magnitude of GRW's spreading rate. If the final value is less than zero, it's set to zero. GRW's spreading rate is zero if no GRW is known and accepted by the EG (i.e. zero is its default value), so PR% will remain at 100% until a GRW is accepted by the EG and will decrease slowly thereafter (i.e. the primitive religion will be inevitably replaced by the GRW).
Civ to Civ Spreading: A random civ is picked in the Known Civs List. If the level of social contact with it is greater than a given number (TBD), continue: Contact Variables in this random civ are changed to represent the contact with the followed religions in the civ being processed. Each Contact Variable in the randomly picked civ is changed to MIN(1,a+b), where a is its current Contact Variable and b is the Acceptance Variable for the same religion in the civ being processed. Example:
"Acceptance Variables" in civ being processed: CHR:1, ISLAM:0, BUD:1
"Contact Variables" in random civ picked: CHR:0, ISLAM:0, BUD:1
New values for Contact Variables in random civ picked: CHR:min(1,0+1)=1, ISLAM:min(1,0+0)=0, BUD:min(1,1+1)=1.
The example shows how the civ being processed exports practiced GRW to a civ with high enough social contact, allowing the latter to have knowledge about them, which will make possible (in the next game turn) for some EGs in it to embrace them if they're accepted. The mechanism also ensures that a civ can only export religions practiced in it, but not those it has contact with, but nobody likes.
The first step makes possible a EG start embracing a GRW. Step two is the EG's internal spreading process leaving the PR in the past. The third step is what makes GRW to spread through the world.
V-3. The "Nationalities" Sub-Model
Nationalism is updated as new_Nationalism=current_Nationalism+C*L, where C is a constant depending on the case (see table) and L is the length in years of the game turn.
________Case______________________________C
EG in a "barbaric province"____________________0.025
EG in a civ with matching nationality______________0.1
EG in a civ with different nationality________0.03*(1/(1+exp(-14+.2*current_Nationalism)))
The rules increase Nationalism as a "natural tendency" as time passes by, making it faster if the tribe has its own civ (govt). But, under the rule of others, Nationalism goes down because it's hard to preserve identity in this conditions. However, it's much easier for tribes with already low Nationalism to lose it than those that already have a great sense of identity and that's why the exponential formula is used.
Nationalities (EGs) with low Nationalism tend to be absorbed by the dominant culture. This happens because the tribe has no sense of itself and easily give up its identity to become something else. This mechanism allows for some tribes to be lost during the game, as it happened with saxons and many other tribes.
The following must be done in each province for each present nationality: If the nationality doesn't match the civ's one, then a dice is rolled to see if some of its population adopts the civ's nationality. The probability of doing so is
P=(1/(1+exp(0.06*current_Nationalism)))*exp(-K*ED)
where ED is the Ethnic Discrimination level the govt imposes and K is a scale parameter. If the dice indicates nationality shift, a population of P*600*L is deducted from the EG's population and added to the EG having the civ's nationality, where L is, again, the length of the game turn in years. This conversion can be up to 3000 persons every 10 years or null, depending on how nationalistic people is and the attitude of the govt regarding other nationalities.
If the population for a given EG falls down to zero, the EG is deleted from the Population Composition (the EG has disappeared at least in this civ).
V-4. The "Cultural Evolution" Sub-Model
This sub-model is the one that changes cultural attributes (characteristics except Nationalism). There are two questions: In which direction the characteristic should go (increase, decrease, stay) and at which speed. Modeling changes in culture is really difficult. It's not easy to answer these two questions (specially the first) in a general way. That's why I grouped characteristics in three categories according to the complexity of the variable:
1) Land Connection
2) Ethnic Tolerance, Importance of Religion and Traditionalism
3) Religious Tolerance, Individualism, Asceticism, Aggressiveness
1) Land Connection
LC can only rise, since it's the love for the province where you live and the notion that the EG has lived there "always". It should increase by around 1 point every 4 years.
2) and 3)
The approach I took is characteristics tend to a value, so in each turn we only move slightly the characteristic in the direction of that value. The magnitude of the change is given by Traditionalism, so the more traditionalistic, the slower the change. The following is made to each characteristic:
new_value=current_value + ((100-Traditionalism)/100)*T*L*(TV_value - current_value)/ABS(TV_value - current_value) (*)
where TV stands for "Tendency Value", T is a time-scale constant and L is the length in years of the game turn. Now the question is what is the TV for each characteristic. The difference between categories 2) and 3) is the way TV is computed.
2) Ethnic Tolerance, Importance of Religion and Traditionalism (TV)
TV_ET=MAX(100-Nationalism, 60)*(1-Aggressiveness/100) (i)
Ethnic Tolerance decreases slowly as a result of the rising Nationalism level and the need to clearly make a difference with other tribes, but Nationalism won't push it below a certain level (60). ET can go lower than this only by the presence of a high Aggressiveness level.
TV_IR= 100-OTL*0.4-ES*3 (ii)
The assumption here is religion's attractiveness is mostly given by the incapability of human beings to understand and control their world. Then, the importance of religion in people's daily lives goes down as science gives us new and useful explanations for our environment and as our society appears to be stable. OTL is the civ's "overall tech level" in a 0-100 range. A measure I hope we can build in the tech model. ES is "Empire's Stability", a variable from the govt model in the range 0-10.
TV_Traditionalism=95-(CivilRights/100)*15 (iii)
This means Traditionalism goes in the 80-95 range depending on the current value of Civil Rights, which is a government model's variable describing how free is people in the 0-100 range. I'm assuming Traditionalism always tends to be high and the differences are from high to very high, depending on freedom. IMO cultures with high freedom give more space to innovation and changes are more acceptable, so Traditionalism tends to be somewhat lower.
3) Religious Tolerance, Individualism, Asceticism, Aggressiveness (TV)
For all these variables religions have something to say and that influence must be merged with other existing factors. The TV will be a mix of the values different factors encourage.
There are 3 factors:
Religion- Through the religion's moral code, religion tries to impose certain values in these characteristics.
Dominant Culture- Majorities tend to impose their customs.
Environment- Effects of economy, govt and others which the people are exposed to in their daily lives. Environment values should be seen as those people would tend to as a result of the environment they live in. In other words, people tend to recreate the environment where they live, making cultural attribute ad hoc to the current environment.
Religion's values are taken from the GRW the EG follows and from the EG's characteristics themselves for PRs. Values majorities imposes are taken from the Majorities Cultural Attributes. (The MCA is created by the govt model using majorities populations. For civs where one EG is above the rest, MCA values are equal to that EG and MCA is a weighted sum of all EGs in the civ for multiethnic civs, where populations are used as weights.) Environment values are the following:
Individualism: You tend to be as individualistic as your environment is. The level of individualism in environment is given by the current economic system (the more capitalistic, the more individualistic) and by the current Civil Rights level, where high CR encourages heterogeneity in population (individualism). The level of individualism by the economic system the govt imposes can be computed as econ_Ind=10*(PP+(10-SP)) where PP=Private Property and SP=Social Policies (both govt model's variables). The civil rights level is given by the Civil Rights govt policy. Assuming economy counts for 70% of total environment influence on individualism, Environment value for Individualism is (econ_Ind*.7+CR*.3).
Aggressiveness: Not yet determined, but it should be a measure of the relationship with neighbors. The more dangerous the world is and the fewer friends abroad, the more aggressive people is as a response in order to survive.
Religious Tolerance: The tendency is to accept other mystic beliefs in the same degree other ideas and lifestyles are accepted. Using the Civil Rights govt policy as a measure of the current level of heterogeneity in population (i.e. level of acceptance of other lifestyles), we make RT_Environment=CR.
Asceticism: Environment (Nature perhaps) imposes the need to survive and gather material goods for it. The assumption is people desire goods and more and more of them until at least a level where you have an accommodated life. We'll define a constant variable for the game: a PCI (per capita income) representing that minimum level of material goods. If the current civ's global average PCI is lower than that, then Asceticism is equal to 100*current_PCI/reference_PCI. If it's higher, Asceticism keeps its current value.
Each of the three factors (religions, Dominant Culture and Environment) push characteristics to given values. The level at which each prevail is another matter. We'll say religion imposes its doctrine depending on the Importance of Religion variable. So, if IR=60, then 60% of the TV for any of the attributes is given by the religion's value. The rest (40%), is influenced by Dominant Culture and Environment and we'll say Environment is responsible for 80% of that:
TV = IR%*religion_value + (1-IR%)*(0.8*environment_value + 0.2*MCA_value) (iv)
Since an EG can support two religions (PR and GRW) each with its own value, the religion_value in the formula is really: PR%*EG_current_value + (1-PR%)*GRW_value.
So, in short, what this sub-model does each turn for each EG is:
1-Increase Land Connection.
2-Compute tendency values for Ethnic Tolerance, Importance of Religion and Traditionalism using (i), (ii) and (iii) respectively.
3-Compute tendency values for Asceticism, Individualism, Aggressiveness and Religious Tolerance using (iv). ("3 factors approach")
4-Apply equation (*) for Ethnic Tolerance, Importance of Religion, Traditionalism, Asceticism, Individualism, Aggressiveness and Religious Tolerance.
Three Comments on the Cultural Evolution sub-model:
1) The model moves characteristics to TV as if TV were an equilibrium point. But, TV values change slowly as things happen, so during the game this "equilibrium point" is moving and characteristics are constantly chasing it.
2) The way TV are computed for characteristics in category 3), i.e. via the three factors approach, allows the model to do at least three interesting things: a)EGs in the same civ will tend to be more homogeneous as time passes by; b)Empires can export their culture to tribes within the empire, like barbaric tribes in central Europe being "romanized" or native americans being "europeanized"; c)The influence of religion on culture will decay overtime.
3) This sub-model is by far the one that needs more computational resources. I guess we can make the model just apply steps 1 and 4 each turn using stored TV, which is really cheap. TV would then be updated less frequently, like twice a century or so.
V-5. The Migrations Sub-Model
NOT YET SOLVED
VI. Final Comments
A lot has been lost from previous attempts to create a social model. This model doesn't have a mixing culture procedure, doesn't model races and race mixing and it's mainly managed at a civ level instead of a province level. Social changes in real life do happen at the local level and then spread, but that kind of modeling is much more expensive in computational resources, so that's why I take care of cultures at the civ level assuming all changes occur simultaneously in each province. As for mixing cultures, it's a growing problem the more nationalities and religions coexist in the same civ and there's no way to have true-to-life effects once the civ-level approach has been adopted.
The current social model just provides a minimum set of interesting effects without making a too complex model. It will be a Clash Team decision if more complexity is mandatory.
I. Introduction
The Clash of Civilizations will be a challenging game not only because of the better AI it will include, but also because the Clash world will be more complex compared to games of this type. Most of models developed in the Clash Project have assumed so far the existence of civ's cultural attributes where they can get inputs from to simulate how different cultures act in different ways and how civilizations can appear to have their own styles and histories based on social effects. It is the main role of this model to produce that necessary social base we're all counting on. This model only describes how people can be characterized through cultural attributes and which mechanisms exist to make attributes change overtime according to the civ's experiences. The effects of culture in the game are supposed to be managed in the rest models. In particular, the Riots Model will take care of major social events such as revolutions, independence attempts, etc.
For the nature of the subject, it's possible for some people to take in the wrong way some opinions about culture and culture evolution. Audiences are encouraged to be open minded when reading the model, because, of course, the only goal here is to replicate what we've seen in history as a general trend and avoid discussions on what's good or bad about cultures.
II. Model Overview
The problems to solve are mainly two: How cultural attributes change as time passes by and how different ethnic groups interact with each other. Modeling the first one is obviously needed, while the importance of the second one needs an explanation: Civs won't interact with each other always like nations, each in its own space. If romans take Egypt, then egyptians end up living under roman rule and they won't necessarily will become roman citizens and support the new regime. Even more, Rome can invade just a part of Egypt, so you can find egyptians in roman provinces and in egyptian civ. In other words, the problem here to solve is the interactions between cultures when more than one occupy the same space (provinces).
The approach I took for the model is this: Each civ can hold several ethnic groups (tribes/nationalities) in it. At a civ level, each of them is described with a set of cultural variables or characteristics. It is assumed that all people with the same nationality have the same characteristics regardless of the place where they live within the civ. This means, FE, that romans in the capital province share the same characteristics that romans living in central Europe provinces. At the province level we only need a dynamic array of the existent nationalities and the populations of each. FE, a roman province in central Europe could be something like this:
Romans-1200
Celts-950000
Franks-1200000
while in a roman province in the greek peninsula would be:
Romans-2600
Greeks-1700000
The number of elements in the arrays change as migrations and other events happen. Populations will be stored on a tile/square basis. Province populations are just the sum of the populations in the tiles forming the province.
The set of ethnic groups (EG) in one civ will be called Population Composition (PC). Each EG in the PC has the following information:
- Nationality (FE, romans)
- PR%: Primitive Religion (share of population who follow the "primitive religion")
- GRW: Great Religion of the World (which GRW is followed)
- GRW's spreading rate
- Religiously Discriminated (dummy variable describing if the EG is discriminated or not by the govt because of religion)
- Ethnically Discriminated (dummy variable describing if the EG is discriminated or not by the govt because of nationality)
- Slavered (dummy variable indicating if the ethnic group is "brutally" slavered, as defined in the govt model when Slavery=2)
- Characteristics (several variables describing the culture)
Each turn what the model does is change the characteristics of each ethnic group in the civ according to some rules explained later and where religion plays a fundamental role. The other important factor is govt. It's a govt decision if any ethnic group is above the rest (FE, romans in the roman empire). In general, the govt will determine a level of ethnic discrimination defining how groups will interact with each other. These two factors, religion and govt, and their role in the model are discussed in the next two sub-sections below in general terms.
It should be clear from what has been said above, that "civ" and "tribe" (or ethnic group) are not synonymous. A player plays a civ, not an ethnic group. The player playing the roman empire controls the roman govt and through it, all what happens in the controlled provinces (the civ), but romans are just one of many EGs in the civ. The romans having a special position respect to other tribes and calling the civ "roman civ" is, above all, a govt decision. The civ may indeed have a nationality of its own. FE, a player might be ruling Yugoslavia (civ's nationality=yugoslavian), while EGs in this civ, although generally referred as yugoslavians, have other nationalities (serbs, croats, etc) and there's no EG above the rest. This example also shows that in this model "nationality" means the tribe you feel you belong to, not something imposed or declared by a govt.
II-1. Modeling Religions
Religions are definitely needed in the game. They have given shape to culture for most of (if not all) history and they have also played huge roles in wars and politics, so they're a cool ingredient to add in Clash. The govt model handles how the Church acts in the political arena. In this section it's explained how religion affects EGs.
A religion will be described by several attributes, many of which are the same that describe a culture. The idea is religions "dictate" to their followers how to behave, so they show the people what values cultural attributes should take. FE, if one religion has an aggressiveness of 70, then all the people who follow that religion will tend to have their cultural attribute "aggressiveness" equal to 70. So, the main role of religions in this model is being culture changers, making cultural attributes alike to religion attributes.
There've been many religions and there're still many of them. Their evolution from their primitive forms to how they're today is a complex matter. In order to avoid the game from keeping track of a lot of religions, their birth, evolution and possible disappearance, I've chosen the following approach: The really important religions are what I've called "the Great Religions of the World". There'll be from 3 to 6 of them in the game (the number will be decided by the team, but that should be the order of magnitude). The model will manage how an EG goes from following primitive religions to following one of the GRW. What makes the difference between a primitive religion (PR) and a GRW is the "exportability" of GRW. PRs are extremely associated with the particular ethnic group that practices it, so it's very difficult for others to embrace the same beliefs and therefore these religions cannot spread. Other peoples can sometimes take elements or even the whole cult, but always adapting it to their own nationality (FE, greek religion passing to romans). If you take ancient egyptian religion or aztec beliefs, it's clear to see how religion has to do with the egyptians and aztecs respectively, but not with humanity in general. GRW go beyond the tribe, so no matter where you come from, it's sufficient to follow what the religion dictates to be part of it and therefore it can spread through the world.
GRW will be described with several attributes shown later. The model won't keep track of PRs' attributes, only GRW.
At the beginning of the game EGs will support a PR with 100%. It's assumed that a PR in one EG is different from one in other EG because of the relationship primitive religions have with nationality. PRs have no name (or can be called after the EG, like "egyptian cult" FE).
The game will randomly give birth to each of the GRW after some thousands years of play. Any EG in any civ will be picked as the starting tribe to embrace it. From that point, the new GRW starts to spread as will be explained. Any EG will be allowed to support only one of the GRW. This is not true in real life because in a group of people different cults can coexist, but when we see the world we notice there's always a very dominant religion and some small groups following other beliefs, so I think the restriction I'm using is good enough and it helps keeping the model simple.
For spreading purposes, in each civ there'll be 2*N dummy variables indicating if each of the GRW is known there and if it has followers, where N is the number of GRW in the game. These will be called "Contact Variables" as in "The civ has/hasn't contact with GRW X" and "Acceptance Variables" as in "There are/aren't EGs accepting the GRW X". FE, in one civ there's contact with christianity and Islam, but not with Buddhism, so the contact variables are CHR-1, ISLAM-1, BUD-0. But only Islam is practiced, so Acceptance Variables are CHR-0, ISLAM-1, BUD-0.
Also for spreading purposes, each civ needs to store its level of "social contact" with the rest civs. For simplicity, that level will be given by the number of trade routes between the two civs. That info will be stored in a list (Known Civs List). Each time a trade route is created or deleted, the lists are updated for involved civs.
II-2. The role of government
Usually a civ's govt is related to a particular EG (a nationality) and it's also common for a govt to be related to a religion. The govt model uses the Ethnic Discrimination (ED) Policy for the first type of discrimination. ED can take the following numbers:
0: No discrimination (uncommon. Think Yugoslavia). When ED=0 the civ is said to be "multiethnic" and the player will be asked to introduce a new name for the civ, like Yugoslavia. This is important, not just flavor. As will be shown, EGs with low Nationalism will tend to lose their identity and adopt the civ's nationality, so a player ruling an empire with many EGs with low Nationalism might choose as a strategy to use a multiethnic civ in order to unite the tribes. This is how the model makes possible the rising of new nationalities like the french from gauls, franks, celts and others. However, this strategy wouldn't work in Yugoslavia, where nationalism in the EGs present was high.
1-9: A nationality is preferred (i.e. only one EG) and all others are ethnically discriminated (Minorities). The civ is named after the EG's nationality (e.g. Roman Civ). This is the most common case. Going from 1 to 8 gives different levels for treatment toward minorities, from respect to other tribes and their customs and the possibility to change nationality to the "dominant" one, to very disrespecting treatment (think nazis).
When a govt chooses to be multiethnic and after a new tribe name is picked, the model merges all cultures to create the characteristics for the new people using relative populations as weights in a weighted sum of the original characteristics. This is no sudden forced culture mixing. The new tribe will have zero population. As the game advances, EGs with low Nationalism will slowly drop their current nationality and become part of the new EG.
The govt can also discriminate via religion. If there's a govt with an intolerant official religion, EGs supporting other beliefs can be considered minorities. This is done in the govt model via the Religious Discrimination variable. Since an EG can follow two religions when the EG variable PR% is greater than 0% and lower than 100%, there's ambiguity in knowing if the EG should be discriminated or not. In that case we'll apply the rule of considering the EG as supporting the GRW alone. This is done because as will be shown, once the EG takes a GRW, it'll inevitably embrace it a 100%, so it's just a matter of time to consider the EG as totally committed with that GRW.
Minorities, then, are all EGs discriminated via nationality and/or religion. The govt discriminates and the whole thing about discrimination has as purpose, here in the social model and in the riots model, to change the attitude of each EG and the things affecting them.
III. Cultural Attributes (Characteristics)
These are the cultural attributes and their role in the game. More can be added given the needs in each of the rest models. All variables are in the 0-100 range.
Ethnic Tolerance: It represents how you see other tribes and what kind of relationship you think exists between your tribe and others. A low ET means you feel your tribe is superior and the rest hardly qualify as human beings. A high ET means you respect other tribes and you're willing to live together with them.
Religious Tolerance: This is the tolerance of other religions (by other I mean the ones not followed by the EG). Low scores mean people feel upset about having to live with people with other cults.
Aggressiveness: This defines how aggressive the culture is. A high score indicates a very warlike culture while a low score indicates peaceful. Leaders of an aggressive culture will find it easier to wage war against their neighbors. A very high score could also give a bonus to the efficiency of troops.
Traditionalism: This represents the people's ties to the past. Its role is force culture to stay as it currently is.
Nationalism: Will be defined as "The notion of being a part of something bigger than your local clan (a nation) and the people's willingness to form that nation and live under one same government of their own". This variable was sometimes known as "provincialism/nationalism". Low Nationalism means you care only about your local clan and you don't see much difference between your tribe and others. A high Nationalism shouldn't be confused with a fanatic or blind nationalism. It's only a rather strong feeling of belonging to a specific tribe and a high opposition to be ruled by other tribe. Nationalism is very low at the beginning of the game (thousands of years before Christ), with some random small differences from tribe to tribe (values in the range 1-10).
Land Connection: This is the people's feeling of belonging to the province they occupy. The love for where they live and the idea that the land they occupy belongs to them. Initial LC will be high because the EG is assumed to have lived there for some time. It'll be used in the Riots Model.
Asceticism: This represents the people's desire for material goods. A high score indicates a lack of interest in material goods, while a low score indicates lots are wanted. It has a role in people's behavior in the economic field.
Importance of Religion: The highest this number, the more influenced the rest characteristics are by religion doctrine. This happens because people want to do what religion says. Importance of Religion will be in general very high at the beginning of the game for all cultures and in general terms it will decrease as time passes by.
Individualism: This is the people's value of self. Its main role is modeling what type of economy people wants. The more individualistic, the more inclined to capitalism and vise versa.
Initial values will be given randomly or using pre-set values defined by us and ad hoc to each historical tribe. The player will decide which option to use.
IV. Attributes of GRW
Each GRW has the following cultural attributes: Religious Tolerance, Aggressiveness, Asceticism and Individualism. They represent the moral code the religion tries to impose on its followers. Religion has no say about the other characteristics described above. The following attributes also exist in GRW:
Sacrifice: The higher this score is the more is required from the followers of the religion. This is money and economic resources the clergy takes away from the people that allows the church to act. The economic model should use this variable to know how much people invest in religious infrastructure.
Countdown: Used to determine when the GRW is born. Its initial value in the game is equal to the date the player chooses to start, doubled (FE, 4000BC implies countdown=2*4000=8000).
Holy Land: The tile in which the religion was born (randomly picked at the moment of religion's birth). This is to model crusades and that sort of things.
After the GRW is born, its moral code remains fixed for the game. Religions have shown to be very reticent to change their doctrines, and although we know some changes have occurred, the assumption made here is still very accurate and keeps the model simple. Sacrifice and Holy Land are also fixed and Countdown stays at zero once the religion is born.
Primitive Religions are supposed to be described with the same variables, but the game won't store that info as with GRW. We'll assume that if you take a EG, the PR associated with it has attributes equal to the EG's cultural attributes. In other words, cultural attributes and religion attributes for PRs are indistinguishable. If you want to know what the chinese PR thinks about individualism, just look at the current individualism score of chinese EG. This is made to avoid keeping track of each of the dozens primitive religions existing and their attributes and to give GRW the importance they deserve, being much more autonomous and homogeneous through out the world. As for sacrifice and holy land, which is info we can't read from the EG attributes, holy land is never needed for PRs and Sacrifice will be assumed to be equal to the average between (100-Individualism) and Asceticism.
V. How the model works
Each turn we do the following:
1. Apply the "GRW birth" sub-model to the game.
2. Apply the "Religion Spreading" sub-model to each civ.
3. Apply the "Nationalities" sub-model to each civ.
4. Apply the "Cultural Evolution" sub-model to each civ.
5. Apply the "Migration" sub-model to each civ.
V-1. The "GRW birth" Sub-Model
The countdown variable for each GRW is checked. If it's zero, do nothing. If it's greater than zero, it's decreased by R*L, where R is a random number between 1 and 3 and L is the length of the game turn in years. If after the decrease it's zero or negative, then set it to zero and give birth to the GRW. This means a random tile is picked anywhere in the world. If no population exist in the tile, another is chosen until a tile with people is found. If there's more than one EG in the tile, one is picked randomly. Contact Variable for the GRW in the civ where the EG lives is turned to 1. The religion's moral code is made equal to the corresponding attributes in the EG. Sacrifice is made equal to the average between (100-Individualism) and Asceticism and Holy Land is made equal to the tile's code number.
V-2. The "Religion Spreading" Sub-Model
For each EG in the civ:
First: If the EG has PR%=100% (no GRW is known and accepted), the model looks at the Contact Variables and for each of them equal to "1", the religion is checked for acceptance by the EG. The EG will accept the GRW if the religion's moral code is not too different from EG's culture. A "distance" between culture and religion's moral code is computed as D=square_root( (C1-R1)^2 + ....+ (Cn-Rn)^2 ), where Ck is the EG's value of the kth cultural attribute and Rk is the religious value for the same attribute. If D is less than a given threshold (TBD), it is accepted. The GRW variable in the EG is turned to the name (or code number) of the GRW accepted and the Acceptance Variable for the GRW in the civ is made equal to 1. A GRW spreading rate is then computed for the EG as (1-Traditionalism)*(K1-D)/K2, where K1 and K2 are scale factors allowing to use D in the calculation so the more similar attributes are, the fastest people will take the religion.
If a GRW is at the govt (i.e. it's official), it has more power to convert people. So, instead of comparing D directly with the threshold, we'll really use D*(1-G*RC%), where G is a dummy being "1" if the GRW being checked happens to be the official one and "0" if not. RC% is the political power of the Religious Class, from the govt model.
Second: If PR% is greater than 0%, then PR% is decreased by the magnitude of GRW's spreading rate. If the final value is less than zero, it's set to zero. GRW's spreading rate is zero if no GRW is known and accepted by the EG (i.e. zero is its default value), so PR% will remain at 100% until a GRW is accepted by the EG and will decrease slowly thereafter (i.e. the primitive religion will be inevitably replaced by the GRW).
Civ to Civ Spreading: A random civ is picked in the Known Civs List. If the level of social contact with it is greater than a given number (TBD), continue: Contact Variables in this random civ are changed to represent the contact with the followed religions in the civ being processed. Each Contact Variable in the randomly picked civ is changed to MIN(1,a+b), where a is its current Contact Variable and b is the Acceptance Variable for the same religion in the civ being processed. Example:
"Acceptance Variables" in civ being processed: CHR:1, ISLAM:0, BUD:1
"Contact Variables" in random civ picked: CHR:0, ISLAM:0, BUD:1
New values for Contact Variables in random civ picked: CHR:min(1,0+1)=1, ISLAM:min(1,0+0)=0, BUD:min(1,1+1)=1.
The example shows how the civ being processed exports practiced GRW to a civ with high enough social contact, allowing the latter to have knowledge about them, which will make possible (in the next game turn) for some EGs in it to embrace them if they're accepted. The mechanism also ensures that a civ can only export religions practiced in it, but not those it has contact with, but nobody likes.
The first step makes possible a EG start embracing a GRW. Step two is the EG's internal spreading process leaving the PR in the past. The third step is what makes GRW to spread through the world.
V-3. The "Nationalities" Sub-Model
Nationalism is updated as new_Nationalism=current_Nationalism+C*L, where C is a constant depending on the case (see table) and L is the length in years of the game turn.
________Case______________________________C
EG in a "barbaric province"____________________0.025
EG in a civ with matching nationality______________0.1
EG in a civ with different nationality________0.03*(1/(1+exp(-14+.2*current_Nationalism)))
The rules increase Nationalism as a "natural tendency" as time passes by, making it faster if the tribe has its own civ (govt). But, under the rule of others, Nationalism goes down because it's hard to preserve identity in this conditions. However, it's much easier for tribes with already low Nationalism to lose it than those that already have a great sense of identity and that's why the exponential formula is used.
Nationalities (EGs) with low Nationalism tend to be absorbed by the dominant culture. This happens because the tribe has no sense of itself and easily give up its identity to become something else. This mechanism allows for some tribes to be lost during the game, as it happened with saxons and many other tribes.
The following must be done in each province for each present nationality: If the nationality doesn't match the civ's one, then a dice is rolled to see if some of its population adopts the civ's nationality. The probability of doing so is
P=(1/(1+exp(0.06*current_Nationalism)))*exp(-K*ED)
where ED is the Ethnic Discrimination level the govt imposes and K is a scale parameter. If the dice indicates nationality shift, a population of P*600*L is deducted from the EG's population and added to the EG having the civ's nationality, where L is, again, the length of the game turn in years. This conversion can be up to 3000 persons every 10 years or null, depending on how nationalistic people is and the attitude of the govt regarding other nationalities.
If the population for a given EG falls down to zero, the EG is deleted from the Population Composition (the EG has disappeared at least in this civ).
V-4. The "Cultural Evolution" Sub-Model
This sub-model is the one that changes cultural attributes (characteristics except Nationalism). There are two questions: In which direction the characteristic should go (increase, decrease, stay) and at which speed. Modeling changes in culture is really difficult. It's not easy to answer these two questions (specially the first) in a general way. That's why I grouped characteristics in three categories according to the complexity of the variable:
1) Land Connection
2) Ethnic Tolerance, Importance of Religion and Traditionalism
3) Religious Tolerance, Individualism, Asceticism, Aggressiveness
1) Land Connection
LC can only rise, since it's the love for the province where you live and the notion that the EG has lived there "always". It should increase by around 1 point every 4 years.
2) and 3)
The approach I took is characteristics tend to a value, so in each turn we only move slightly the characteristic in the direction of that value. The magnitude of the change is given by Traditionalism, so the more traditionalistic, the slower the change. The following is made to each characteristic:
new_value=current_value + ((100-Traditionalism)/100)*T*L*(TV_value - current_value)/ABS(TV_value - current_value) (*)
where TV stands for "Tendency Value", T is a time-scale constant and L is the length in years of the game turn. Now the question is what is the TV for each characteristic. The difference between categories 2) and 3) is the way TV is computed.
2) Ethnic Tolerance, Importance of Religion and Traditionalism (TV)
TV_ET=MAX(100-Nationalism, 60)*(1-Aggressiveness/100) (i)
Ethnic Tolerance decreases slowly as a result of the rising Nationalism level and the need to clearly make a difference with other tribes, but Nationalism won't push it below a certain level (60). ET can go lower than this only by the presence of a high Aggressiveness level.
TV_IR= 100-OTL*0.4-ES*3 (ii)
The assumption here is religion's attractiveness is mostly given by the incapability of human beings to understand and control their world. Then, the importance of religion in people's daily lives goes down as science gives us new and useful explanations for our environment and as our society appears to be stable. OTL is the civ's "overall tech level" in a 0-100 range. A measure I hope we can build in the tech model. ES is "Empire's Stability", a variable from the govt model in the range 0-10.
TV_Traditionalism=95-(CivilRights/100)*15 (iii)
This means Traditionalism goes in the 80-95 range depending on the current value of Civil Rights, which is a government model's variable describing how free is people in the 0-100 range. I'm assuming Traditionalism always tends to be high and the differences are from high to very high, depending on freedom. IMO cultures with high freedom give more space to innovation and changes are more acceptable, so Traditionalism tends to be somewhat lower.
3) Religious Tolerance, Individualism, Asceticism, Aggressiveness (TV)
For all these variables religions have something to say and that influence must be merged with other existing factors. The TV will be a mix of the values different factors encourage.
There are 3 factors:
Religion- Through the religion's moral code, religion tries to impose certain values in these characteristics.
Dominant Culture- Majorities tend to impose their customs.
Environment- Effects of economy, govt and others which the people are exposed to in their daily lives. Environment values should be seen as those people would tend to as a result of the environment they live in. In other words, people tend to recreate the environment where they live, making cultural attribute ad hoc to the current environment.
Religion's values are taken from the GRW the EG follows and from the EG's characteristics themselves for PRs. Values majorities imposes are taken from the Majorities Cultural Attributes. (The MCA is created by the govt model using majorities populations. For civs where one EG is above the rest, MCA values are equal to that EG and MCA is a weighted sum of all EGs in the civ for multiethnic civs, where populations are used as weights.) Environment values are the following:
Individualism: You tend to be as individualistic as your environment is. The level of individualism in environment is given by the current economic system (the more capitalistic, the more individualistic) and by the current Civil Rights level, where high CR encourages heterogeneity in population (individualism). The level of individualism by the economic system the govt imposes can be computed as econ_Ind=10*(PP+(10-SP)) where PP=Private Property and SP=Social Policies (both govt model's variables). The civil rights level is given by the Civil Rights govt policy. Assuming economy counts for 70% of total environment influence on individualism, Environment value for Individualism is (econ_Ind*.7+CR*.3).
Aggressiveness: Not yet determined, but it should be a measure of the relationship with neighbors. The more dangerous the world is and the fewer friends abroad, the more aggressive people is as a response in order to survive.
Religious Tolerance: The tendency is to accept other mystic beliefs in the same degree other ideas and lifestyles are accepted. Using the Civil Rights govt policy as a measure of the current level of heterogeneity in population (i.e. level of acceptance of other lifestyles), we make RT_Environment=CR.
Asceticism: Environment (Nature perhaps) imposes the need to survive and gather material goods for it. The assumption is people desire goods and more and more of them until at least a level where you have an accommodated life. We'll define a constant variable for the game: a PCI (per capita income) representing that minimum level of material goods. If the current civ's global average PCI is lower than that, then Asceticism is equal to 100*current_PCI/reference_PCI. If it's higher, Asceticism keeps its current value.
Each of the three factors (religions, Dominant Culture and Environment) push characteristics to given values. The level at which each prevail is another matter. We'll say religion imposes its doctrine depending on the Importance of Religion variable. So, if IR=60, then 60% of the TV for any of the attributes is given by the religion's value. The rest (40%), is influenced by Dominant Culture and Environment and we'll say Environment is responsible for 80% of that:
TV = IR%*religion_value + (1-IR%)*(0.8*environment_value + 0.2*MCA_value) (iv)
Since an EG can support two religions (PR and GRW) each with its own value, the religion_value in the formula is really: PR%*EG_current_value + (1-PR%)*GRW_value.
So, in short, what this sub-model does each turn for each EG is:
1-Increase Land Connection.
2-Compute tendency values for Ethnic Tolerance, Importance of Religion and Traditionalism using (i), (ii) and (iii) respectively.
3-Compute tendency values for Asceticism, Individualism, Aggressiveness and Religious Tolerance using (iv). ("3 factors approach")
4-Apply equation (*) for Ethnic Tolerance, Importance of Religion, Traditionalism, Asceticism, Individualism, Aggressiveness and Religious Tolerance.
Three Comments on the Cultural Evolution sub-model:
1) The model moves characteristics to TV as if TV were an equilibrium point. But, TV values change slowly as things happen, so during the game this "equilibrium point" is moving and characteristics are constantly chasing it.
2) The way TV are computed for characteristics in category 3), i.e. via the three factors approach, allows the model to do at least three interesting things: a)EGs in the same civ will tend to be more homogeneous as time passes by; b)Empires can export their culture to tribes within the empire, like barbaric tribes in central Europe being "romanized" or native americans being "europeanized"; c)The influence of religion on culture will decay overtime.
3) This sub-model is by far the one that needs more computational resources. I guess we can make the model just apply steps 1 and 4 each turn using stored TV, which is really cheap. TV would then be updated less frequently, like twice a century or so.
V-5. The Migrations Sub-Model
NOT YET SOLVED
VI. Final Comments
A lot has been lost from previous attempts to create a social model. This model doesn't have a mixing culture procedure, doesn't model races and race mixing and it's mainly managed at a civ level instead of a province level. Social changes in real life do happen at the local level and then spread, but that kind of modeling is much more expensive in computational resources, so that's why I take care of cultures at the civ level assuming all changes occur simultaneously in each province. As for mixing cultures, it's a growing problem the more nationalities and religions coexist in the same civ and there's no way to have true-to-life effects once the civ-level approach has been adopted.
The current social model just provides a minimum set of interesting effects without making a too complex model. It will be a Clash Team decision if more complexity is mandatory.
Comment