What defines a Bowline? - structure, characteristics, topology

[Dan_Lehman]

I guess that before we can get down to defining a Bowline we have to reach some agreement on how a Bowline should be dressed and set.  I have climbed on the bowline for a number of years and I would NEVER have dreamt of climbing on a knot so shoddily dressed.  I have however seen knots at the waterfront where the collar has been hugely distended and the half hitch has rotated to nothing more than a turn.  I had always assumed that these (to me failed) examples were either the toll of time or just shoddy knotcraft.

You can see rather loose collars in photos of bowlines
in sailing/yachting magazines, and at the waterfront, as you
note; I have posted images here of bowlines that have capsized
into a sort of pile-hitch noose structure (with yet much curve
in the SPart) --something I've seen to such frequency to make
me wonder if the result is intended!?

Different materials give different results (along with varied forces);
rockclimbers' kernmantle ropes won't hold such a loose shape
(and might not hold any unextended bowline) very long,
if at all --the stiff rope wants to "unwind."

Of course this is not what the Bowline would ever be dressed like, ...  [rdl : Oh, I won't go this far!]

What I try to point out is that whichever way you try to dress the bowline,
when it is in use, the nipping turn takes a spiral form resembling that of the Gleipnir,
...

Hmmm, I don't see the turNip in the Gleipnir getting so far as
a spiral --which is dangerously close to losing grip.  Rather, it only
goes towards a turn from the set *hitch* geometry.

The Bowline always takes that form under load,
...
 Below I inserted a photo of a Bowline that ... was exposed to heavy stress ...

Yes, and I see more *hitch*/=>*turn* than *spiral* in this, still.


--dl*
====

[Dan_Lehman]

YES ! This is a detail of a picture of a properly dressed and tightened bowline.

The loadings of the standing end, the eye leg of the standing part, and the eye leg of the bight happen to be equal. 
   Make an educated guess : Is it a commn bowline, or an Eskimo bowline ?
   Derek, give up... There is no way to distinguish if this is a detail of common bowline or a Eskimo bowline !
Why then the one is a bowline, and the other a non-bowline loop ?

There are three *ends*, Xarax : why do you offer only two choices?!

And the point is --re "games"-- that you here have shifted the issue
from What distinguishes <any_eyeknot> from <other_eyeknots>?
--i.p., what characterizes *a* "bowline"?--
to the more general, basic, and profound question
What defines an eyeknot? !!

For the structure you present, with the equal angles of incidence
among *ends* (need a better term for "limbs" of knots in general),
behaves w/o difference so long as the loading is preserved.
(In practice, though, were an eye extant and figuring in the loading,
increased loading would show itself with a gradual change in the
angles and associated tensions of the three *ends*.))

Incidentally, I fancy one of the "Janus" structures partly with
the somewhat whimsical thought that, if one just ensures that
collars go around both an eyeleg & the SPart --no matter the
order--, one is okay with the result either way (for those who
might find making the discrimination a major challenge
--the UNinterested knotting folks) !

And re the profound issue manifest in this example, it begs
the question of using "eye" by me in "eyeknot" --for there
might in fact be no such structure.  (My example builds from
using a bowline to fore corner cleats of a small barge,
of cleating the line (to improve its steerage), and then some
accident severs the bight between cleats along the barge
--the *knot* all this time is closely surveyed and never
changes in discernible quality, yet clearly with the severed
rope there is no (longer an) eye.)

In the case of building definitions, we should look to some
established *canonical* form and not to exotic cases;
I submit that the canonical form of an(y) *eyeknot* is a knot of
a single Piece_of_Flexible_Material with 4 *ends* of which One stands
in tensional opposition to another Two, and the 4th is untensioned
(and, yes, it matters that the tensioned one flows into one of
the tensioned 2, not to the untensioned 4th (which would be
e.g. a *bight hitch* if so).)


--dl*
====

--dl*
====

[xarax]

There are three *ends*, Xarax : why do you offer only two choices?!

 Because I didn't want to offer to Derek Smith and you the easy satisfaction to be able to eliminate the third possibility... (The third choice can be eliminated by the first sight, you know...)

need a better term for "limbs" of knots in general

  I agree on this. Suggest something."Limb" sounds good...

behaves w/o difference so long as the loading is preserved.

  That was exactly my main point. A theoretical counter-example, in a special, yet no "limit" case. This is the case we really have a Janus-like situation, where the two "faces" of the two bowlines structures, that of the common bowline, and that of the Eskimo bowline, are equivalent and indistinguishable.

if one just ensures that collars go around both an eyeleg & the SPart --no matter the
order--, one is okay with the result either way

  That was also one of my points, so the common bowline and the Eskimo bowline are indistinguishable, even in their security/holding aspects. However, I think that it is better to U-turn the first collar around the most tensioned "limb", which, in most cases, is the standing end outside the eye. That is because the first collar bears the most of the tension forces, and we do not want this collar to be deformed, due to a deformation of the segment of the rope around which the collar is U-turned, and which the collar pulls towards the nipping loop.. I might be mistaken in this, though...You could examine this in more detail.

In the case of building definitions, we should look to someestablished *canonical* form and not to exotic cases

  Yes, I agree, and I have said the same when DerekSmith offered his example of the 0%/100% loaded bowline, which is indistinguishable, indeed, from the Sheet bend. ( and it is the ONLY case where this happens,,,) However, my 120 degrees counter-example is not an exotic case, in the sense that it is not a case on the "limit" ; it is a case on the "middle", albeit a very special one.

[xarax]

@ Derek Smith, with playful intentions...

[DDK]

. . . There is no way to distinguish if this is a detail of common bowline or a Eskimo bowline ! . . . 

Unless I'm missing something, you must mean a common cowboy bowline as I do not think this can be a common bowline.  Terrific use of symmetry to help capture some thought-provoking points.

DDK

[xarax]

you must mean a common cowboy bowline as I do not think this can be a common bowline. 

   Derek Smith got it, too...I should have left longer tails ... I use the term "common bowline", to denote both, the left and the right handed bowlines, in contrast to the term "Eskimo bowline" ( an Eskimo bowline can also be tied in a left and a right handed form). I would prefer a pair of antonyms, for those two similar types of end-of-line loops, perhaps (+) bowline and (-) bowline, cis- and trans- bowline  , or something like that...The pair "bowline-antibowline" might also be acceptable, if it is interpreted like the "particle-antiparticle", or "cyclone-anticyclone" pair, and not as acceptance - denial of some essential bowline property, present in the bowline and absent in the anti-bowline.

[Dan_Lehman]

The pair "bowline-antibowline" might also be acceptable, if it is interpreted like the "particle-antiparticle",
 or "cyclone-anticyclone" pair, ...

It is precisely this 2nd that is my motivator; and the point of
determination is the side of the nipping loop from which the
tail is returned (enters) --on the SPart's side (i.e., where SPart
crosses itself) or opposite.

(And beyond this I have "false bowline" though "pseudo-bowline"
might be better (unless I determine separate tasks, each in need).)

 

[xarax]

the point of determination is the side of the nipping loop from which the tail is returned (enters) --on the SPart's side (i.e., where SPart crosses itself) or opposite.

   I, too, agree on this..

I have "false bowline" though "pseudo-bowline" might be better

...but not on this, of course !   I wish to retain the generality in the definition of a bowline, independently of the particular side of the nipping loop the tails enters it for the first time, or exits from it right afterwards. A general definition would be a definition for the (+) bowline, as well as for the (-) bowline, the Eskimo thing. If there is a miracle in the bowline mechanism, it takes place in the common bowline AND in the Eskimo bowline, and this has to do with how perfectly the spectacular anti-slippage action of the nipping loop is combined with the mechanical-capstan advantage of the collar, to generate this miraculous end-of-the-line fixed loop.
   The reason I do not use the term "antibowline" is just this : one can very easily, almost unnoticeably, jumb from an antonymic definition, denoting simple opposition, to a definition denoting negation ( of the "bowline property" ). The counter-example of the 120 degrees bowline, where we are not able to distinguish, by a local inspection, if we have a common bowline or an Eskimo bowline, is not an"extreme" case, a case "on the limit": On the contrary, it is a middle-ground case, a case "on the middle",  that should have convinced you, by now...Allow me to repeat/sing a previous post :

 I believe that, from time immemorial, people tied bowlines in both ways, just preferred the common bowline, when the eye legs were close to be parallel ( in loops tied around small diameter objects, where the knot s nub is far from the object ), and preferred  the Eskimo bowline, when the eye legs were close to be aligned ( in loops tied around large diameter objects, where the knot s nub is close to the object). There is no essential difference : the collar, an extension of the eye leg of the bight, helps the nipping loop secure the tail, and makes a U turn around the line that is more aligned with the eye leg of the bight. If the loop is elongated, because it is tied around a small diameter object, the collar makes a U turn around the segment of the standing part outside the bight. If, on the other hand, the loop is round, because it encircles a large diameter object, the eye leg of the bight is more aligned with the segment of the standing part inside the bight, so it is natural to the collar to make a U turn around this segment of the standing part, i.e. around the eye leg of the standing part.

[[Inkanyezi] gone]

I think I am beginning to get xarax's vision here. The most extreme case of the Bowline is the most common one, where the two legs of the eye are almost parallel and pointing away from the standing part, and it is in this form the turNip is evident. When legs are wider apart, the nipping turn takes another form, which is the same as in the Eskimo Bowline.

Now I have tried to collapse bowlines that are tied around very wide objects, and they will not collapse, neither will the turn take the turNip form, but it resembles the Sheet Bend form somewhat, although differently loaded.

In order to make the knot collapse completely, I tried easing the collar a lot, but still it failed to collapse, even in shock cord. It winds the collar around the knot and still holds. But when the collar is snugly worked from start, it only goes so far as to allow the standing part to exit the knot almost straight from the nipping turn. (Which of course is a round turn, rather than a spiral, we already had a name for that.)

I don't know whether the discussion can be taken much further, and I still think of it as more a theoretical than a practical issue. I learned a bit more about one of our most used knots, and it did not change the world.

[DerekSmith]

In light of the comments by Xarax and Inkanezi, I have just completed a number of trials, static loading and dynamic loading, well dressed and set Bowlines (#1010) in a number of different cords ranging from 0.8mm Spectra right up to 18mm cored braid.

The objective was to determine if the #1010 Half hitch Component underwent transformation into the Turn Component as suggested when the Bowline is loaded.

The result is that indeed IT DOES...

The more 'elastic' the cord, the lower was the tension needed to rotate the Half Hitch Component (HhC) into a Turn Component (TC), and in the most elastic Polyester braid, the Turn Component even disengaged contact of the coil and started to morph into an Open Spiral Component, while at the other end of the spectrum, very low stretch materials were virtually at breaking point before any noticeable morphing could be detected (and then only slight).

Of Note; once the bight collar has been elongated (pulled through the nip) by the tension induced rotation, it remained extended in a form approaching the 'sloppily dressed' state I commented on earlier.

Consequently, it is necessary for me to amend my proposed definition a third time, to incorporate the morphing from Half Hitch Component to Turn Component when the knot is subjected to high loads in relatively extensible cords.

A Bowline is defined by Ashley in #1010.
----------------------
It is a loop knot comprising two interlocked components -  a Bight Loop Component and a Half Hitch Component
Under conditions of high load in extensible cordage, the Half Hitch Component will morph open to form the Turn Component and may even show signs of further morphing into the Spiral Component.

[NOTES - with the exception of the 12mm dynamic Kern-mantle safety rope and the 18mm cored braid (their tensile strength exceeded the capacity of my test facility), the test cords were extended until they failed with the Bowline - this was assumed to be ca 50% of line strength.  In tests, the cords were then subjected to ca 30%  loading by taking the knot to 60% of the extension exhibited at failure.]

Derek

[Dan_Lehman]

the point of determination is the side of the nipping loop from which the tail is returned (enters) --on the SPart's side (i.e., where SPart crosses itself) or opposite.

   I, too, agree on this..

But a problem I run into is that many such antibowlines have a strong
tendency to exhibit less a **loop** than a **spiral** --YMMV per knot,
but in some, it's quite hard to have the part that should be a **loop**
staying something approaching a plane, crossing points proximate,
even if not touching.  AND, moreover, it might be argued that such
a **spiral** base is DESIRED/expected.

And I really want a classification that doesn't hinge on loading variance
(why I favor calling a "noose" things that show the structure's SPart
really playing no **knotting** role but just being hitched to (e.g.,
the venerable two half-hitches & midshipman's hitch
--these, to me, are not, in their entirety, **knots**, but **compound
structures** which contain a knot (the hitch to their SParts)).

I have "false bowline" though "pseudo-bowline" might be better

...but not on this, of course !  

You might if you understand my meaning.  "Read more than 1 in 30 words..." !
I indicated that these were other terms and for other uses.
I.p., for knots in which there appears to be (or is) a nipping loop
but in which the continuation does not go into an eye leg (immediately),
but e.g. forms a collar around the eye legs, and THEN feeds into one
of them.  So, one has the apparent structure of a loop but not quite
--"false"/"pseudo".  One might expect nipping & strength characteristics
to be similar, and knots of this sort can have your "proper collar" and
be easily untied; in the slippery HMPE cordage they will likely NOT
slip (collapse the eye as shown for the dbl.bwl  by Brion Toss's video).

I wish to retain the generality in the definition of a bowline,
independently of the particular side of the nipping loop the tails enters it for the first time,
or exits from it right afterwards.

I too want generality (if it indeed can work --see my doubts re **spiral**),
and more so than you I thought, re "proper collar",
except in the enboldened utterance above where I think you must
imply accepting the *improper* simple wrap I would take,
such as in the Myrtle --which serves a collaring role,
but has the tail exiting through the loop opposite to #1010.

A general definition would be a definition for the (+) bowline,
 as well as for the (-) bowline, the Eskimo thing

I think I'm coming to see the latter as a boundary-crossing knot
--in the canonical form of SPart-vs-eye_legs (not equilateralism),
setting the collar tight will produce a core nipping structure as
much a crossing knot as a turNip. and I'm thinking that
it's best not to try to push this one way or another, just as with
that carrick loop which can also have either core nipper.

NB: We have followed the OP's topic of trying to articulate the essence
of "a bowilne "; but this might be a biased start --to pic from one
reasonable set of knots this one and then ... :  we might have
started with a broad set of knots (albeit those thought to be like
the bowline and then sought to characterize & group them,
maybe rejecting some as belonging.  And in such an exercise,
I submit that one might see JUST the turNip as the essence,
and then look to group according to what knots preserve that.

 I believe that, from time immemorial, people tied bowlines in both ways,

I clipped a clothing(?) ad that featured a yacht in which the two lines
tied to the sail were . . . tied in both ways !  And I surmised ignorance
more than some equanimity or love of variety (or even pressure of some
circumstance) --conjecture.


--dl*
====

[xarax]

 many such antibowlines have a strong tendency to exhibit less a **loop** than a **spiral**
  it's quite hard to have the part that should be a **loop*staying something approaching a plane, crossing points proximate,
even if not touching.

   We might expect that a helical (helix, not spiral) nipping structure, where the two points - when/where there should be only one - are not very far from each other, is more effective, as a nipping structure, that when we have a "proper" plane nipping loop, and one crossing point, but a "twisted" nipping loop, with limbs in a pronounced angled/elbow configuration. In the helical case, we have a slightly less nipping power from the maximum, but in the "twisted" case, we have much of the nipping power wasted by friction, and much of the tensile forces dissipated by the oblique loading of the two limbs.
   So, we do not know where we have a more effective nipping structure, and where we have a less effective one, i.e. we can not correlate the helical structure with a supposedly less of a bowline "anti- bowline" property, and the plane structure with a "genuine bowline...Why we should prefer a slightly twisted plane nipping loop, as in a (+) bowline, than a slightly open helical nipping structure, as in the (-) "anti-bowline" ? No, the fact that the nipping structure is open or closed, is not a measure of the effectiveness of the basic bowline structure, so it can not be an element of "discrimination" between the two forms of bowlines.

  And I really want a classification that doesn't hinge on loading variance...

  Me too. More than that, I want a classification that does not takes into account even the global differences of the two forms - because we have such a strong local symmetry, that is always existent, but it is revealed most clearly in the 120 degrees "equilateral" ( strictly speaking, equiangular) case, the Janus (+) bowline / Janus (-) bowline, common bowline / Eskimo bowline structure.

  I indicated that these were other terms and for other uses, I.p., for knots in which there appears to be (or is) a nipping loop, but in which the continuation does not go into an eye leg (immediately),  but e.g. forms a collar around the eye legs, and THEN feeds into one of them.  So, one has the apparent structure of a loop but not quite --"false"/"pseudo".  One might expect nipping & strength characteristics to be similar, and knots of this sort can have your "proper collar" and
be easily untied;

   They do not have a "proper "nipping loop !   Lets agree that the nipping loop should feed the standing end and the eye leg of the standing part immediately/directly, and not through St Peter s Cathedral, the Great Wall or the Great Barrier reef... Why reserve any compound "bowline" name for such a compound "not-bowline" knot ? We should not use the term "bowline", for any of those knots !

the *improper* simple wrap...such as in the Myrtle --which serves a collaring role, but has the tail exiting through the loop opposite to #1010.

   We have enough problems with the "proper" or not "collars", we should not try - at least for the time being - to solve even more, by accepting "proper" and not "collaring roles"   ! A not-proper collar-role-playing, collar-like structure, can be even more effective than a proper collar, indeed, but if I accept it as a collar, I open the Pandora s box, in the sense described by knot4u...
   (And the tail is not "exiting through the loop opposite to #1010 "  I refuse to describe every simple fact of ropes and knots structures by an Ashley number !! The tail is not exiting through the loop from the opposite direction it followed when it was entering into it for the first time, as it should, if the collar was a "proper" bowline one.)

  I think I'm coming to see the [Eskimo thing] as a boundary-crossing knot...setting the collar tight will produce a core nipping structure as much a crossing knot as a turNip.,  and I'm thinking that it's best not to try to push this one way or another

  So, either you throw out with the beautiful baby Eskimo, with the dirty water of the crossing knot-based structures, like the ugly Karash thing, or you accept them both as bowlines. Back to the square one !  

...I submit that one might see JUST the turNip as the essence,and then look to group according to what knots preserve that.

  If so, you should be prepared to deal with many more "bowlines" than the average knot tyer, bowline user, or man in the street / harbour would like to, believe me...What the purpose of a definition that only a few people will accept, fewer still will use, and only a handful will understand ?  
   (Heraclitus of Ephesus : People must "follow the common (hepesthai tō ksunō)" and not live having "their own judgement (phonēsis)". )

[Dan_Lehman]

I indicated that these were other terms and for other uses,
i.p., for knots in which there appears to be (or is) a nipping loop, but in which the
continuation does not go into an eye leg (immediately),
...
 

They do not have a "proper "nipping loop ! ...
 Why reserve any compound "bowline" name for such a compound "not-bowline" knot ?

Because they appear to be such (as is the case for some
such names in biology, e.g.) --and will take scrutiny to
conclude "Aha, it is NOT (truly) ... !".  (And some folks like
Derek will have some satisfaction in seeing the semblance
acknowledged this much, at least.)

(But these are still just sketches of thought at the inchoate
formation of a possible nomenclature, yet to be brought into
even some "beta" stage of application.)

the *improper* simple wrap...such as in the Myrtle --which serves a collaring role,
but has the tail exiting through the loop opposite to #1010.

   We have enough problems with the "proper" or not "collars",
we should not try - at least for the time being - to solve even more,
by accepting "proper" and not "collaring roles"   !
...

Well, of course, I actually don't care about the collar, as I'm looking
at the eye knots that are based on the turNip core, which use
and stabilize that --the Gleipnir has no collar at all, but to make
an eye knot one needs to somehow effect the stabilization of the
turNip without the benefit of opposed tails brought through it;
I'm happy with whatever works.  ("The" bowline just happens to
come along as one solution, among the many --many unknown.)

  I think I'm coming to see the [Eskimo thing] as a boundary-crossing knot...setting the collar tight will produce a core nipping structure as much a crossing knot as a turNip.,  and I'm thinking that it's best not to try to push this one way or another

  So, either you throw out with the beautiful baby Eskimo, with the dirty water of the crossing knot-based structures, like the ugly Karash thing, or you accept them both as bowlines. Back to the square one !  

No, I shrug at the *knots* that can take forms that confound my
classification, that span my artificial boundaries --the duckbilled
platypuses of knots, if you will.

...I submit that one might see JUST the turNip as the essence,and then look to group according to what knots preserve that.

  If so, you should be prepared to deal with many more "bowlines" than the average knot tyer, bowline user, or man in the street / harbour would like to, believe me...What the purpose of a definition that only a few people will accept, fewer still will use, and only a handful will understand ?  

That cannot be avoided --the "deal(ing) with..."--, for the knots are
there (to be discovered, to be ignored), no matter the naming.  And
a slew of (non-"bowline") names isn't going to make it any simpler.
(And when you realize what quite simple things the occasional rope
user & knotter can confuse, you hardly want to limit your own doings
to whatever lower common denominator might exist (you might doubt
there is even any!) !!)


--dl*
====

[xarax]

   I have just did what it seems that you suggest : I forgot the bowline solution, I tied a planar nipping loop, and I have thought of ways to stabilize that structure by using the available free end, the "tail", for that purpose, and that purpose only :  to prevent the planar nipping loop turn into a helical nipping loop, and finally into a straight line, when loaded. In 15 minutes I came up with 15 solutions... that achieved that purpose and that purpose only, and only as a collateral by-product, those solutions managed to secured the tail itself as well...( And this happened even if they were no meant/designed to do this, it happened just because of the necessarily convoluted structures formed by the tail, that were subsequently tightened by themselves, "locked" on the nipping loop s limbs, and turned into safe hitches...)
   What I mean is this : If you forget that the main purpose in the bowline is to secure the tail - not to stabilize the coil of the nipping loop - and that this purpose is achieved by constricting the tail twice ( at least ) into a nipping loop, and after, and after only, this main task has been achieved, and as a collateral by-product, and as a collateral by-product only, this nipping loop happens to be stabilized as well, by the mere presence of the formed structure by the tail, the one (at least) "proper" collar, if you forget this, you open the Pandora s box, or a can of anti-knot4u worms !
   The main purpose of the stone-age man that discovered the bowline, was to secure the tail on the standing part, to form a stable end-of-line loop, to tie his wife with... He figured out to form a nipping loop on that part, so he could pass the tail through this nipping loop, and consrict and secure the tail there, by the nipping action of the two loaded limbs of this loop. Once going through the nipping loop was proven- by trial and error- to be not enough, so he tried to do the same thing twice... He could, at this time, have tied a common nipping loop, or a Eskimo nipping loop, or a Myrtle, we do not know...But then, the "proper" collar solution survived the evolution pressures by what ( incidentally, as a collateral by-product ) this structure  achieved, in an almost perfect way, by just "being there" : the stabilization of the nipping loop. The "proper" collar solved the problem of the tail, that was, and still is, the main problem of any end-of-line loop, and then, miraculously, a historical accident took place : the stone-age man, (and at least one of his descendants... ), realized that this most simple structure was preventing the nipping loop to be turned into a helix, and, eventually, into a straight line...This first ingenious brave man died, some of his descendants happen to survive, and many of them still live : but they forgot the purpose of their ancestor, and turned the whole thing upside down! They think that the stabilization of the nipping loop is the main task of the bowline, and, as this task is achieved, by whatever sufficiently convoluted structure, the tail will be secured as well - will happen to be secured. Even if this is not meant to happen, if the initial motive is long forgotten, it will happen :  the tail will be secured, as a collateral by-product of the convoluted structure it forms.
   Wrong, this view of our ancestor s tool...Not wrong because it leads to wrong new knots ( it might well lead to better, if only more complex end-of-line loops ), but because it leads to wrong definitions of the old ones, the knots in the bowline family. I tell you, forget the main purpose of the bowline, confuse the main purpose of the nipping loop - that has to do with the tail, and not with itself ! - forget the presence of the "proper" collar that is but the simplest structure that achieves to pass the tail through the nipping loop twice, and, at the same time, happen to stabilize this nipping loop as well by its mere presence, and then you enter into a vast realm of possible end-of-line loops...We can say that they are inspired by the bowline, but it is far fetched to say that they are bowlines. I would be glad to get rid of this pesky Karash thing and all the possible crossing-knot based loops, but keep the baby Eskimo alive within the tube, and here comes this descendant, who wishes to open the knot4u can of worms...and force the few remaining readers start running towards the tree swing...

[DDK]

. . .  We have followed the OP's topic of trying to articulate the essence
of "a bowilne "; but this might be a biased start --to pic from one
reasonable set of knots this one and then ... :  we might have
started with a broad set of knots (albeit those thought to be like
the bowline and then sought to characterize & group them,
maybe rejecting some as belonging.  And in such an exercise,
I submit that one might see JUST the turNip as the essence,
and then look to group according to what knots preserve that.  . . .
--dl*
====

These comments strike a chord as I wonder how different our discussion would be if we were attempting to describe a family of knots which we were looking to label as the TURNIP LOOPS.  In this case, is it not likely that a significant variety of structure and function would be more easily expected and accepted?

If we label a family as BOWLINE LOOPS, do we not because of the notoriety of the bowline have some narrower expectations of structure and function?  I believe the bowline is very widely known as a rescue loop, not exclusively, but prominently.   Do we want to label knots as bowlines which would / could / should never be used as a rescue loop?  I would have no issues with a member of the TURNIP LOOP family not being a rescue loop.  I feel somewhat differently about members of the BOWLINE LOOP family not being able to function at some level as rescue loops (security, ease of tying, etc.).

In my opinion, the use of the name of an archetype in the naming of a family of knots has in some cases more specific or restrictive implications.

DDK