I would like to start a discussion on posssible causes for the saeculum. I'd like to start with what Strauss and Howe suggested as a cause of the cycle in Generations. They used the concept of generational constellation in their discussion of causes. The generational constellation is easiest to understand in terms of an idealized set of generations, each of which is as long as a 22-year-long phase of life. Assume at time zero generation type D starts being born; at year 22 generation type C starts being born; at year 44 generational type B starts being born and at year 66, generation type A starts being born. Now imagine New Year's day in year 88. Assume everyone before year zero is dead. How old will members of each generation be?

Members of generation type A will be age 0-21 years, and will exactly fill the youth phase of life. Generation type B will be age 22-43 and will exactly fill the rising adult phase of life. Everyone who is in the mature adult phase of life (age 44-65) will belong to the type C generation and everyone over age 65 will belong to a type D generation and occupy the elder phase of life.

If time is moved up 22 years, each generation moves up one notch: A to rising adult (age 22-43); B to mature adult (age 44-65);and C to elderhood (age 65-87). Generation D will have passed from the scene, and a new generation of the same type will have been born to replace them. This generation type D will fully occupy the youth phase of life. Strauss and Howe call the specific times, when generations match up with phases of life, aligned generational constellations. When an aligned constellation occurs, a new generation starts being born. Two to five years after this, a new turning starts. It is the periodic alignment of generations with phases of life that acts as the pacemaker for the saeculum.

The question is how well does this work?

Of course, real generations are not typically exactly 22 years long. One can still identify aligned constellations by calculating the percentage of each phase of life that is filled by the proper type of generation. For example, consider 1767, the first year of the Compromise generation. This generation is of the adaptive type, making generation D in the earlier example adaptive. The table lists the other generations.

Generational constellation on New Year's Day 1767

Youth = 1745-1766--> Civic Gen = 1742-1766 . Youth 100% Filled by Civics
Adult = 1723-1744--> Reactive .= 1724-1741 . Adult 82% Filled by Reactives
Mature=1701-1722--> Idealist . = 1701-1723 . Maturity 100% Filled by Idealists
Elder = 1679-1700--> Adaptive, = 1674-1700 . Elderhood 100% Filled by Adaptives

Each phase of life is completely filled by the correct generation, except Rising Adult. The reactive Liberty generation was only 18 years long and cannot fill more than 82% of a 22-year phase of life. Thus, each phase of life in Table 8 is maximally filled, making 1767 a perfectly aligned constellation.

Most generational constellations are not so well aligned. For example, the 1618 constellation (1618 is the first year of the reactive Cavalier generation) shows 100% youth, 41% rising adult, 64% maturity and 50% elderhood occupation by the appropriate generations--not a very good alignment. Yet the alignment of the very same generations with phase of life is much better in the years just before 1618. In 1609, the alignment was best: 95%-95%-100%-91%. The year 1609 can be considered as the optimally-aligned generational constellation for the start of the Cavalier generation. It can be considered as the predicted date for the start of the Cavalier generation based on the generational constellation. In this particular case, the actual start of the Cavalier generation (determined from biographical information) was nine years after the predicted date. The start of the Compromise generation in 1767 was exactly on schedule according to the constellation model, while the 1943 start of the Boomer generation was four years before the constellation prediction. These differences between the actual values and the expected values based on the model are called model residuals.

The residuals represent that portion of real-world behavior not explained by the model. A good model will show residuals that are randomly distributed around zero. Any pattern to residuals implies that some important factor has not been captured by the model. The Figure shows a plot of the residuals for the generational constellation model.



The residuals are not randomly scattered. Rather, they show a highly significant downward trend with regression coefficient R of 0.82, indicating 99.999% statistical significance. Also they average 2.6 years rather which is more than 10% of a generational length away from zero. When Strauss and Howe developed this model, they did not have the first six generations in Table 3. If the residuals obtained from these generations are left out the average value of the residual falls to just 0.7. It is clear that the constellation model was developed because it fit the observable data on average quite well. But when the structure of the residuals are examined it is clear that the model does not fit the data well.

This is interesting, but you refer to a table. Can you please post the link?

Thanks.

Originally Posted by The Wonkette

This is interesting, but you refer to a table. Can you please post the link?

Thanks.

Table 3 is simply the list of turnings and generations, it's right here at this site. I put in a link.

Originally Posted by Mike Alexander '59

The residuals are not randomly scattered. Rather, they show a highly significant downward trend with regression coefficient R of 0.82, indicating 99.999% statistical significance... But when the structure of the residuals are examined it is clear that the model does not fit the data well.

This interesting analysis makes the premise of this website null and void, doesn't it?

Originally Posted by David Kaiser

I admit our future lies with the Millennials.

Short answer... No, not at all. :wink:

Originally Posted by Devil's Advocate

Originally Posted by Mike Alexander '59

The residuals are not randomly scattered. Rather, they show a highly significant downward trend with regression coefficient R of 0.82, indicating 99.999% statistical significance... But when the structure of the residuals are examined it is clear that the model does not fit the data well.

This interesting analysis makes the premise of this website null and void, doesn't it?

Not exactly. It doesn't negate the existence of the saeculum or the generations S&H found. It merely casts doubt on the idea that the generations interacting with phases of life cause the cycle.

This is not to say generations and phases of life aren't involved, merely that they are not the whole story.

Originally Posted by Mike Alexander '59

Originally Posted by Devil's Advocate

Originally Posted by Mike Alexander '59

The residuals are not randomly scattered. Rather, they show a highly significant downward trend with regression coefficient R of 0.82, indicating 99.999% statistical significance... But when the structure of the residuals are examined it is clear that the model does not fit the data well.

This interesting analysis makes the premise of this website null and void, doesn't it?

Not exactly. It doesn't negate the existence of the saeculum or the generations S&H found. It merely casts doubt on the idea that the generations interacting with phases of life cause the cycle.

Your analysis certainly invalidates the manner by which S&H were discerning generational boundaries over the last 500 years. Or, the analysis shows a definitive trend narrowing the span of a generation from ~22 to less than 18 years (of which William Strauss has recently dismissed as a mere "conceit of the present day").

Given lengthening life-spans (ie, an energetic, so-called, GI living long and heathy enough to proudly and triumphantly usher in a so-called 3T), one, based on your analysis, has to seriously question the relevance of the latter conclusion.

Otherwise, shouldn't one of sober mind seriously consider "tossing out" the S&H generational theory altogether?

Originally Posted by Devil's Advocate

Originally Posted by Mike Alexander '59

Originally Posted by Devil's Advocate

Originally Posted by Mike Alexander '59

The residuals are not randomly scattered. Rather, they show a highly significant downward trend with regression coefficient R of 0.82, indicating 99.999% statistical significance... But when the structure of the residuals are examined it is clear that the model does not fit the data well.

This interesting analysis makes the premise of this website null and void, doesn't it?

Not exactly. It doesn't negate the existence of the saeculum or the generations S&H found. It merely casts doubt on the idea that the generations interacting with phases of life cause the cycle.

Your analysis certainly invalidates the manner by which S&H were discerning generational boundaries over the last 500 years.

How so? S&H claim they determined the generations by examining cohort biographies. If their 22-year phase-of-life model was explicitly used to obtain the generations, then they would have averaged 22 years long before 1700. They don't.

The phase of life model is an attempt to explain the generations they did find. It doesn't do a good job of explanation because if you use it to forecast generations, they do average 22 years long and the model soon runs way ahead of the actual generations.

Or, the analysis shows a definitive trend narrowing the span of a generation from ~22 to less than 18 years (of which William Strauss has recently dismissed as a mere "conceit of the present day").

No, the analysis shows a decline in average generational length from about 27 years to 20 years.

Given lengthening life-spans (i.e., an energetic, so-called, GI living long and healthy enough to proudly and triumphantly usher in a so-called 3T), one, based on your analysis, has to seriously question the relevance of the latter conclusion.

Otherwise, shouldn't one of sober mind seriously consider "tossing out" the S&H generational theory altogether?

You are confusing (as many here do) the theory S&H give for why generations occurs with the observation they made that such generations exist. This is like throwing out the idea that planets exist because the Ptolemaic system didn't work to describe their movement.

Originally Posted by Mike Alexander '59

Originally Posted by Devil's Advocate

Your analysis certainly invalidates the manner by which S&H were discerning generational boundaries over the last 500 years.

How so? S&H claim they determined the generations by examining cohort biographies. If their 22-year phase-of-life model was explicitly used to obtain the generations, then they would have averaged 22 years long before 1700. They don't.

I posted Neil Howe's methodology not long ago. They used a combination of bios and historical events to discern generational boundaries. Furthermore, as Howe noted, the further back in history one goes, the less, not to mention less reliable, information there is to work with. This is obviously problematic to a methodology that relies heavily on accurate information. Otherwise, one is left wide open to swallowing myths and distortions.

Or, the analysis shows a definitive trend narrowing the span of a generation from ~22 to less than 18 years (of which William Strauss has recently dismissed as a mere "conceit of the present day").

No, the analysis shows a decline in average generational length from about 27 years to 20 years.

Their analysis ranged, in just 150 years, from 30 years (Transcendental) to just 18 years (Boom).

That's too much slop over too short a period of time (CW anomaly notwithstanding), even if one is relying on averages. Your analysis shows a steady declining trend, of which even the huge burp of the S&H anomaly failed to slow.

Methinks S&H fell headlong into the very trap of a "conceit of the present day" for one reason or another. But that's just my opinion.

I think I'll stay out of this argument, but I would like to add a comment.

Models are always that: models! Lilke any other model, actors outside the model can cause the model to be wildly inaccurate at times. Just look at the similarly complex system models we use to predict weather.

I think the saecular model does a pretty good job of describing the cyclical aspects of history. I also believe it's incomplete. The one glaring error is the drift to shorter turnings and generations. We seem to need a theory that explains the gradient - a theory that is simply missing - for now, at least.

Mike's analyses are a very good starting point for identifying the underlying trend, but trends are only aggregated data. We still need the AH-HA moment that adds the explanation of why the trend exists. No, I'm not volunteering one of my own.

Originally Posted by Devil's Advocate

Originally Posted by Mike Alexander '59

Originally Posted by Devil's Advocate

Your analysis certainly invalidates the manner by which S&H were discerning generational boundaries over the last 500 years.

How so? S&H claim they determined the generations by examining cohort biographies. If their 22-year phase-of-life model was explicitly used to obtain the generations, then they would have averaged 22 years long before 1700. They don't.

I posted Neil Howe's methodology not long ago. They used a combination of bios and historical events to discern generational boundaries. Furthermore, as Howe noted, the further back in history one goes, the less, not to mention less reliable, information there is to work with. This is obviously problematic to a methodology that relies heavily on accurate information. Otherwise, one is left wide open to swallowing myths and distortions.

You stated that my analysis invalidates the methods S&H used to determine generations. My analysis invalidates their concept of the generational constellation as the cause of the change in generations. Unless they used this concept to determine their generations what you said isn't correct. I replied that if they had used the constellation concept to determine the generations they would be 22 years long prior to 1700. Your most recent response is doesn't address my point at all. Do you understand it?

Originally Posted by Marx & Lennon

I think the saecular model does a pretty good job of describing the cyclical aspects of history.

What do you mean by saecular model? Do you mean the idea that generational constellations by themselves cause turning changes? In that case I would disagree.

Or do you mean the idea that history can sometimes be usefully characterized in terms of a repeating series of generational archetypes?

Mike,

Some very elegant analysis as usual. However, does your conclusion somewhat rest on an implicit assumption that everyone in a a particular generation is equally important, fungible, to history, and that sheer numbers, in the form of the ideal 100% representation, is necessary to cause the saeculum?

Could it be that a certain sub-population (e.g. elites, clergy, trend-setters, rock stars, in-the- right-place-at-the-right-time) that makes up less than the ideal 100% population for the correct positioning could still have the characteristic of its respective generation at the respective turning to influence to the degree necessary to cause the saeculum?

Afterall, even Presidential election 'landslides' are often a result of only 2 or 3 percentage points one way or another.

Originally Posted by Strauss and Howe

Originally Posted by Mike Alexander '59

My analysis invalidates their concept of the generational constellation as the cause of the change in generations.

Q. In a nutshell, what drives the cycle?

A. History creates generations, and generations create history. The cycle draws forward energy from each generation's need to redefine the social role of each new phase of life it enters. And it draws circular energy from each generation's tendency to fill perceived gaps and to correct (indeed, overcorrect) the excesses of its elders. The powerful nurturing and “shadow” relationships between two-apart generations are especially important. The alternation between underprotection and overprotection of children is also key.

Unless they used this concept to determine their generations what you said isn't correct. I replied that if they had used the constellation concept to determine the generations they would be 22 years long prior to 1700. Your most recent response is doesn't address my point at all. Do you understand it?

I understand your point, I just don't understand the relevance of a history where a relatively recent cycle lasts 110 years (the New World cycle) and today's Millennium cycle is merely 72 years long.

I mean, shoot, let's just determine we're in a perpetual "Crisis" and be done with any historical meaning of it all (which I have always assumed was the #1 S&H point).

Originally Posted by Devil's Advocate

Their analysis ranged, in just 150 years, from 30 years (Transcendental) to just 18 years (Boom).

That's too much slop over too short a period of time (CW anomaly notwithstanding), even if one is relying on averages. Your analysis shows a steady declining trend, of which even the huge burp of the S&H anomaly failed to slow.

Actual the decline in generation length is not steadily, but rather a step change.


This plot shows turning lengths from McGuinnes before 1435 and Strauss and Howe afterward. The McGuinness turnings average 26.3 years in length while the S&H generations between 1433 and 1821 average 25.9 years in length. Clearly generations/turnings averaged 26 years over the 1147-1821 period. The regression line show no significant trend in length over the entire 650+ year period. Generations after 1822 are much shorter, averaging only 20 years in length. These generations also show no significant trend.

The entire trend reflects a step change in length in the early 19th century.

Mike,

It seems that you're letting Youth = 0 - 21, Young Adult = 22 - 43, Midlife = 44 - 65, and Elderhood = 66+... letting each phase be the same length, 22 years.

But, of course, S&H did not quite have it that way; their Youth phase was one year shorter (0 - 20), and the other phases all ended one year earlier as well so we had one 21 year phase and the other phases being 22 years... would that difference (which makes the phases of life slightly less symmetric) change things any?

Originally Posted by Devil's Advocate

I understand your point, I just don't understand the relevance of a history where a relatively recent cycle lasts 110 years (the New World cycle) and today's Millennium cycle is merely 72 years long.

That's what they found. And when I take a different approach, using different methods, I find a similar result. The question I am asking on this thread is why does this happen? What sort of mechanism(s) can produce these observable shifts in length? As my plot shows the change in length was both large and happened pretty rapidly. I interpret this large change as a change in fundamental cause, so I think of the pre-1820 saeculum and the post-1820 saeculum as having different basic causes.

Originally Posted by Seminomad

Mike,

It seems that you're letting Youth = 0 - 21, Young Adult = 22 - 43, Midlife = 44 - 65, and Elderhood = 66+... letting each phase be the same length, 22 years.

But, of course, S&H did not quite have it that way; their Youth phase was one year shorter (0 - 20), and the other phases all ended one year earlier as well so we had one 21 year phase and the other phases being 22 years... would that difference (which makes the phases of life slightly less symmetric) change things any?

S&H had youth 0-21 in Generations. From a theoretical POV, If the phases of life are not the same length then it is impossible in even the ideal case for a prefectly aligned constellation to occur.

Practically it would change nothing. The trending residual problem can be fixed easily enough by simply assuming longer phases of life, say 26 years before 1820 and 20 years afterward. There are problems with this approach which I will get to as soon as I write it up.

Originally Posted by salsabob

Mike,

Some very elegant analysis as usual. However, does your conclusion somewhat rest on an implicit assumption that everyone in a a particular generation is equally important, fungible, to history, and that sheer numbers, in the form of the ideal 100% representation, is necessary to cause the saeculum?

Could it be that a certain sub-population (e.g. elites, clergy, trend-setters, rock stars, in-the- right-place-at-the-right-time) that makes up less than the ideal 100% population for the correct positioning could still have the characteristic of its respective generation at the respective turning to influence to the degree necessary to cause the saeculum?

Afterall, even Presidential election 'landslides' are often a result of only 2 or 3 percentage points one way or another.

Certainly, a subgroup can cause the cycle. In fact it probably does. But how does this change anything? They will still experience human lifecycles, defined by S&H as four phases of life. As I noted above, if we change the definition of phases of life we can make a phase of life model fit the data.

What we would be doing is making generation length an adjustable parameter, which we fit to the data to obtain some form of minimized residual. For example, linear regression, which you may be familiar with is a method of curve fitting in which the sum of the square of the residuals is minimized. Mini-max fitting minimizes the sum of the absolute value of the residuals.

The constellation model mathematically comes down to generation length = phase of life. So one simply sets phase of life length equal to average generation length and you have completed a mini-max type curve fitting optimization.

A saeculum model uisng L = 26 for t < 1820 and L = 20 for t > 1820 will fit the data quite well. Indeed it can be used for prediction as well:

Next crisis = 2009 = 1929+80, or 2006 = 1946+60 or 2004 = 1964+40 or 2004 = 1984+20. These average to 2005, that value S&H give in T4T, and value favored by many here.

It is an easy matter to come up with a formulation that fits the data. The question is whether it is physically realistic. Is there evidence to support the idea that phases of life were longer in the past? A physically unrealistic model cannot be correct.

Originally Posted by Mike Alexander '59

Originally Posted by Marx & Lennon

I think the saecular model does a pretty good job of describing the cyclical aspects of history.

What do you mean by saecular model? Do you mean the idea that generational constellations by themselves cause turning changes? In that case I would disagree.

Or do you mean the idea that history can sometimes be usefully characterized in terms of a repeating series of generational archetypes?

I'm more in line with your second comment. I have a suspicion that the correltaion between turnings and generations is just that, a correlation. I have a harder time with cause and effect. It may be that the two are mutually reinforcing, in a positive feedback sort of way. Obviously, it's much harder to analyze than physical systems with much better defined properties.

It may be that humans are predisposed to change, and modernity has allowed this inherent human virtue to emerge. That a pattern has developed is not surprising. That S&H did a good job of describing the pattern is their strongest claim to fame.

While I agree that we will be moving into a new turning soon, if not already at the beginning of one, the timing precision that you are analyzing has never been convincing to me. I tend to see the chaos effect making that impractical. That doesn't invalidate the basic concept, though.

Originally Posted by Mike Alexander '59

Originally Posted by Devil's Advocate

I understand your point, I just don't understand the relevance of a history where a relatively recent cycle lasts 110 years (the New World cycle) and today's Millennium cycle is merely 72 years long.

That's what they found. And when I take a different approach, using different methods, I find a similar result. The question I am asking on this thread is why does this happen? What sort of mechanism(s) can produce these observable shifts in length? As my plot shows the change in length was both large and happened pretty rapidly. I interpret this large change as a change in fundamental cause, so I think of the pre-1820 saeculum and the post-1820 saeculum as having different basic causes.

Here's some pure speculation, for discussion purposes only:

Assume that the cycle is also dependent on hysteresis, and that hystersis is determined by the rate at which concepts are communicated through society. As the world has grown more advanced in this area: better methods of printing, better roads, better physical conveyances, and finally - virtually instantaneous communications, hysteresis has declined.

This is certainly simplistic, but works as a basic hypothesis. If applied to the Civil War anomoly, the argument could be made that the emergence at that time of railroads and later of the telegraph produced a singular shift of substantial magnitude and created a disruption. Of course, it could also be hogwash.

Originally Posted by Marx & Lennon

I have a suspicion that the correlation between turnings and generations is just that, a correlation.

I'm not sure there is an empirical correlation between S&H turnings and generations. If you look at the tight correspondence between their generations (being born) and their turnings, it is obvious that one is derived from the other. Nothing in social science correlates that well.

Although S&H give textual descriptions and definitions of turnings, they provide no way to determine them. They do talk about how they went about obtaining generations in Generations. Since they published their generations before the turnings, it makes sense that they discovered the generations first and came up with the turning dates later based on their generation. I don't think they did an independent "turning hunt" and came up with the set they published.

I have constructing an independent set of turnings. It correlates in a statistically significant fashion with the S&H turnings although thy certainly do not overlay them. Not only that, but the turnings also correlate with Dave McGuiness's pre-1435 turnings. Now there is no way Dave's pre-1435 turnings can be dependent on S&H because S&H say turnings don't exist before 1435. So Dave's pre-1435 turnings are independent of S&H's. S&H's turnings are necessarily independent of both Dave's and my work because they published first.

I used a different method than S&H did (and probably Dave too). This makes my work independent of S&H and probably Dave. Yet all three independent lines of work show agreement suggesting that Strauss and Howe saw something real when they looked at generations.

Obviously, it's much harder to analyze than physical systems with much better defined properties.

Not at all. Unless you believe in intelligent design or vitalism, issues such as these should be amenable to the same sort of analysis and study as any non-experimental natural science.

Originally Posted by Mike Alexander '59

Originally Posted by Marx & Lennon

... Obviously, it's much harder to analyze [history] than physical systems with much better defined properties.

Not at all. Unless you believe in intelligent design or vitalism, issues such as these should be amenable to the same sort of analysis and study as any non-experimental natural science.

<PHILOSOPHICAL_BABBLE>
I had to think about this before I responded, and I'll agree that analyze is the wrong word to describe what I mean. In fact, you can analyze virtually anything, and gather something of value from it. I meant to say that technical analysis will not, in and of itself, tell us much about the underlying mechanisms that produce the patterns the analysis reveals. That requires insight and the wisdom to recognize it when you see it.

I once used a sig block that said, in essense, that knowledge, intelligence and wisdom are not the same, though they complement one another well. I still believe that, though the intangibles: insight and inspiration, should have been included, too. The quest for knowledge is always a good thing, well ... usually anyhow, and there's no shortage of brainpower on this forum.

I'll credit S&H for inspiration. We're all here because of it. Their insight may be validated or not in the next few years. I'm withholding judgement on 'wisdom'. We may not live long enough to answer that one. In any case we still have plenty of time to argue about both.
</PHILOSOPHICAL_BABBLE>

I declare you all to be way off here.

Imagine you wanted to determine why a car travelled from one place to another, and you looked at the answer purely in terms of the movement of wheels and gasoline explosions moving pistons and friction and gravity and so on. Is this why the car moved from one point to another, or did it happen because I chose to drive it there?

The saeculum happens because the world changes and societies, groups of people, need to change to better fit themselves to their new environments. The different turnings, the differences between generations, these are the mechanism by which society is able to change, not the cause of it.

In a completely stable and unchanging environment, there would be no saeculum because there would be no need for it. Each generation would be like the one before it, and you'd have something like a High, some milder form of it, which would go on forever.

Why didn't the High of the 1950's go on forever? Because the world had been changing for the previous decades and centuries and american society (looking at this from an american point of view) was very badly suited for its environment in a number of ways. For example, half the population, the female half, was supposed to spend their lives sitting around at home baking cookies and vacuuming. This was left over from a time when most of the population were farmers and women were having 6 or 8 kids per family and they'd be spending their days caring for all these children and doing the large amounts of work necessary to keep up a home.

Shift to a technological society, and suddenly there is no need to have so many kids, no pigs and chickens to care for, most of the household chores are done by machines and the few kids are in school all day, and now women have a role in society which makes no sense. In order to fit its new environment, society needs to change to allow women to fill the roles which they are now capable of and suited for. But people don't want to change, they are highly resistant to it. How do they change? The answer is, the saeculum.

Looking at it from this perspective, it is easy to see why the saeculum has speeded up and the generations have gotten shorter in the last century or two. The world has been changing at an ever increasing rate, as technological progress has exploded across the planet. Society can barely keep up with all the changes, and the saeculum is seemingly being driven forward as fast as it can go.

There's a lot more one can understand about the saeculum if you understand the actual purpose behind it. It is not just a wheel rolling along throughout the centuries for no reason.

Originally Posted by Starkk

The saeculum happens because the world changes and societies, groups of people, need to change to better fit themselves to their new environments. The different turnings, the differences between generations, these are the mechanism by which society is able to change, not the cause of it.

Why should change occur in a cyclical fashion?

Originally Posted by Mike Alexander '59

Originally Posted by Starkk

The saeculum happens because the world changes and societies, groups of people, need to change to better fit themselves to their new environments. The different turnings, the differences between generations, these are the mechanism by which society is able to change, not the cause of it.

Why should change occur in a cyclical fashion?

Perhaps the best answer, the only one that makes sense, is...

... Why not?