Salt in the Gospel of Mark

My purpose here is to take a literary approach to the gospel of Mark, and argue for a new exegesis of his “salt” sayings (Mk. 9:48-50). I will not be trying to relate this to the historical Jesus, nor will I be considering its implication for Christian faith. The goal here is do try to determine what the original author of the gospel of Mark meant to say in these lines.

My exegesis is: Where ‘their worm does not die and the fire is not quenched’ everyone will indeed be salted (purified, made acceptable to God) by fire. The salt (that makes sacrifices acceptable to God) is good. Now, if salt (of the covenant, that makes things acceptable to God) were to become salt-less (unacceptable to God) in nature, in whom will it be seasoned (preserved, renewed)? Have (my) salt (that now makes one acceptable to God) in you, and be at peace with one another.

Here, on my reading Mark has crafted a section involving two interlocking metaphors that refer to Leviticus. There is the “salt of the covenant” (Leviticus 2:13), and there is also the salt that must be present in cultic sacrifices to make them acceptable to God (Leviticus 2:11,13) which leads to the salt that must be in metaphorical sacrifices following Christ, to make them acceptable to God.

Text and notes

The relevant pericope is given below. Note that after John’s question there is one continual speech by Jesus. Any break in the text of Mark here is artificial, introduced by looking ahead to see what Matthew and Luke did with this material.

Mark 9:38-50 – “John said to him, ‘Teacher, we saw someone (who was (or ‘is’) not following us) driving out demons in your name, [EN TW ONOMATI SOU] and (because he was not following us) we were preventing him (or ‘we stopped him’).’ But Jesus said, ‘Do not stop him; for [GAR] there is no one who will do a miracle in the name of me [EPI TW ONOMATI MOU] and will soon afterwards be able to speak evil of me. Indeed, [GAR] anyone who is not against us is for us. Indeed, [GAR] if anyone gives you a cup of water to drink because you are in the name of Christ, [EN ONOMATI OTI CRISTOU ESTE] then in truth I tell you, he will most certainly not lose his reward. But anyone who causes one of these little ones that have faith to stumble, it is better for him if he is constrained like a mill-donkey around his neck [EI PERIKEITAI MULOS ONIKOS PERI TON TRACHLON AUTON] and is thrown into the sea. And if your hand causes you to stumble, cut it off. It is better for you to enter into the life maimed than with two hands to go into [Gehenna], (into the unquenchable fire). And if your foot causes you to stumble, cut it off. It is better for you to enter into the life crippled than to have two feet and be thrown into [Gehenna]. And if your eye causes you to stumble, pluck it out. It is better for you to enter the new rule [BASILEIA] of God with one eye than to have two eyes and be thrown into [Gehenna]. Where ‘their worm does not die and the fire is not quenched' everyone will indeed be salted with fire. [OPOU...PAS GAR PURI ALIAQHSETAI] The salt is good. [KALON TO hALAS] Now, if salt were to become salt-less in nature, [EAN DE TO hALAS ANALON GENHTAI] in whom will it be seasoned? [EN TINI AUTO ARTUSETE] Have (my) salt [hALA] in yourselves, and be at peace with each other."

I’ve placed a period after the 3rd “Gehenna”. The ‘worm and the fire’ and ‘salted by fire’ are linked by GAR.

I translated EN TINI as “in whom”, based on the repetition starting at the end of the previous pericope, of “in my name”, “in your name”, “in my name”, “in the name of Christ”.

For the last salt, Mark writes hALA instead of hALAS. Why is unclear, but it seems that it is somehow to be understood differently than the hALAS salt. Danker1 has that it is “probably a back-formation from hALAT- ” So hALA could be sort of a “case-less” salt. It could then intentionally be read either as “Have salt in you” (accusative) or “Have my salt, or the salt of the one of whom I speak, in you” (genitive). 

The rest of these notes rely on France’s commentary3.

The text of v.38 has complex variants, but they do not have a large impact on the meaning. “Is/was not following us” may occur twice or only once, and if it occurs only once it may occur in either position.

Some texts repeat EPI TW ONOMATI MOU in v.41 rather than EN ONOMATI OTI

Some texts add EIS EME in v.42 (faith “in me”).

EIS TO PUR TO ASBESTON, ‘”into the unquenchable fire”’ may be in both v.43 and v.45 or in neither.

Some witnesses repeat the text of v.48 in v. 44 and v.46, but a wide range of witnesses omits the repetition.

Alternate readings in v.49 are “All sacrifices will be salted with fire” and “Everyone will be salted with fire and all sacrifices salted with salt”.  However “Everyone will be salted with fire” is regarded as the most probable original reading.

Background Information

In Greek mythology salt is called a divine substance.19 Plato, in the Timaeus, describes it as especially dear to the gods.18 One possible explanation for this association is that salt appears to have eternal properties. You can dissolve salt in water, and then recover it. Numbers 18:19 describes “an eternal covenant of salt”. The idea of a “covenant of salt” between God and his people in Hebrew scripture begins in Genesis with Abram’s victory at “the Valley of Siddim (now the Salt Sea)” (Genesis 14:3, New Jerusalem). It is explicitly mentioned first in Leviticus 2, “You will put salt in every cereal offering that you offer, and you will not fail to put the salt of the covenant of your God on your cereal offering”. Ancient Greeks, Egyptians, and Romans also followed the practice of putting salt in sacrifices.6

Hospitality was an important virtue, tied to the idea of covenants with God. So salt, which sealed covenants with God was also a symbol of hospitality. The Arab expression 'There is salt between us' reflects the idea that to eat another man's salt creates a sacred bond between host and guest.13 A Russian tradition of hospitality, extending into modern times, khlebosol’stv is literally “regaling with bread (khleb) and salt (sol)”.16

Kosher meat was salted to remove the blood. And in Arabic acceptable food is ‘halal’. This is similar to the word for salt in both ancient Egyptian (hal) and in Greek (hALAS). The Hebrew word for “the Law”, “halachah”, comes from the word “halach” meaning “the way to walk or to go”. This seems strikingly similar to the word for “salt” in languages that Hebrew interacted with.17 Lot’s wife was turned into a pillar of salt. And, salt was rubbed on newborns (Ezekiel 16:4). This practice of protecting newborns with salt continued into the early modern period in Europe.6

St. Augustine (Conf., I. 1, c. xi) mentions salt as a religious symbol. “I was regularly signed with the cross and seasoned with his salt (et condiebar eius sale), even from the womb of my mother.” This is significant, since it shows that Christians in late antiquity had a notion of “The salt of Christ”. There are also other references to salt in connection with baptism from the early church.11 In the Clementine Homilies, there are reference to Christians, probably related to the Ebionites, that practice a Eucharist involving bread and salt.12 The Homilies themselves were probably produced by these groups in the early 3rd century, but may refer to practices that were older. These views lingered around the area east of the Jordan, and then were absorbed into Islam in the 7th century.14 

Salt is still found in modern religious symbolism. On Shabbat, modern Jews dip bread (or Challot) in salt before eating them.9 In addition salt plays a role in Passover observance.10 Salt also still has a liturgical role in the modern Catholic Church.11 And, of interest, there is a small group known as the Mandaeans, that number only about 20,000 in modern day Iraq and Iran that claim origins older than Islam and Christianity. Various speculations have placed their origin with early Jewish Christians or perhaps with the Jewish sect associated with John the Baptist.15 They have rituals for purification of all food, utensils and equipment for cooking, with the exception of salt, which is considered pure.

All this symbolism based on salt seems a little strange to modern ears, but salt was a very valuable commodity in the ancient world. The ancient Chinese kept a government monopoly on it, and their economy was largely based on iron and salt.6 Salt was also central to the economy of the Celtic peoples north of Rome. The Roman name for these people “Galli” (or Gauls) comes from the word for salt in Greek “hALAS” and Egyptian “hal”.6 Kurlansky6 also writes that the “the Romans…called a man in love ‘salax’, in a salted state, which is the origin of the word ‘salacious’”. Actually it might be better to just note that the root word “sal” means salt, and “in lust” might be better than “in love” here. “Salarium”, “money for salt”, from which we get our word “salary”, was pay for Roman soldiers. Typical pay for Roman workers was also in the form of bread, wine, salt, and olives. One of the oldest of the great Roman roads, the Via Salaria, was the “Salt Road”, and most Italian cities were founded near saltworks, starting with Rome in the hills behind the saltworks at the mouth of the Tiber.6 While the Romans did not have a government monopoly on salt, they did carefully regulate its price for political purposes.


I believe the author Mark has crafted a puzzle or riddle that we are supposed to solve. This raises a couple of special problems for exegesis here. First, if I am correct, Mark is engaging in eisegesis using the text of Leviticus. So here we need to be careful that the eisegesis is what the author Mark intended, and not ours. This is compounded by the fact, that in order to solve a riddle, one needs to “try on” various solutions in order to see which one fits best. This is a highly inductive process, and again, for this reason, we need to be careful to avoid doing our own eisegesis on the text. My solution is to carefully control the introduction of subjectivity by using Bayesian analysis. While this is not a typical approach to New Testament studies, I believe the exercise is useful. If it is not possible to construct an argument for a given point in terms of Bayesian statistics, it becomes apparent that the point in question is highly subjective, and is likely something that is being read in to the text, rather than something that can be arrived at by exegesis on the text itself.

Bayesian analyst represents one “school” of statistical thought. Another school of thought defines probability in terms of frequency. The main difference is that those that hold the “frequentist” point of view see probability as an external reality. Those in the Bayesian camp see probability as something that is related to a degree of internal subjective certainty. For example, if the only information we have is that there are 6 horses in a race, then Bayesian probability would assert that the probability of each horse winning is 1/6th. The frequentist would envision an infinite sequence of identical races from which we have not yet drawn a sample. The frequentist wants to estimate the real fraction of races that would be won by each horse in that infinite set of races. The Bayesian does not envision estimating anything here, 1/6th is the correct answer for the given information. If our information set changes, say we learn something about the past racing history of the horses, their starting positions, or whether or not one has a broken leg then that would change the correct answer for the Bayesian. Then both the frequentist and the Bayesian will be interested in the result of the first race, but they will treat the information differently. The Bayesian will use the information to update his subjective probability, or his justified certainty, regarding the outcome of the next race. The frequentist will treat the race as a sample of the hypothetical infinite pool of identical races. For the frequentist, this information is the used to reduce the confidence interval on the estimate of objective (external) probability.

In general, the controversy between the schools is very abstract and involves how Bayesian analysis can be grounded in mathematical axioms, and the philosophical foundation of mathematics. In most practical cases, either the Bayesian or the frequency formulation can be used and they give the same results. However there are a few extra things that can be done with Bayesian analysis and no one denies that these extra things are useful. Jaynes’s book4 is an excellent reference for Bayesian analysis. It not only describes practical implementation, and important issues like unbiased prior probabilities, but also describes the philosophical issues, and controversies, and develops Bayesian logic as an extension of formal logic.

While Bayesian analysis is subjective, it need not be subjective in the bad sense, where everyone will get their own answers. The correct answer in Bayesian analysis depends on the information that is considered. As long as individuals agree on what information is to be considered, then Bayesian probability will yield an exact result that is the same for everyone. A legitimate criticism of the result can never take the form, “You don’t have enough information to know that”, because in Bayesian probability if there is enough information to define the question, then there is enough information to assign starting probabilities to the possible answers. For example if we say that “wugs” are either “snarky” or “snicky” and then ask what the probability is that the first wug we meet will be snicky, we can say the probability is 50% by enumeration of the possibilities. Saying, “But we don’t know anything about wugs.” is not a legitimate criticism of this result. 50% is the correct answer for the given information.

On the other hand, a quite legitimate criticism of a Bayesian result is that important information, that is known, was not considered that would effect the outcome. For example, if we know it is daylight, and that snicky wugs only come out at night, then we have omitted important information that will change our answer. Our first answer was not wrong; it was correct for its information set. And, with our new insight into the nocturnal behavior of snicky wugs, we now have a new correct probability for our new information set. As an analogy, what axioms are to deductive logic, the information set is to Bayesian analysis. A deductive argument may be completely sound, but it is only as good at approximation of "truth" as its axioms. In Bayesian analysis if the math is done correctly, the answer is only as good of an approximation of reality as the information fed into it.

One of the most important tools in Bayesian analysis is Bayes’s formula. It gives a precise method for updating a probability assessment as new information is incorporated.

Let h = prior estimate of probability of hypothesis.
Let h|e = probability of hypothesis given the new event.
Let e|h = probability of the event given the hypothesis.
Let e = total probability of the event on any hypothesis.

Then h|e = (h * e|h) / e.

Here is an example. Suppose the apriori probability that it is raining outside is 30%.
Also, let’s say that if it is not raining there is a 50% chance that the cat is outside. But if it is raining then there is only a 10% chance that the cat is outside. Now, if we determine that the cat is outside, how does this change our assessment of the probability that it is raining?

h = apriori probability of rain = 30%
h|e = rain given cat outside = unknown
e|h = cat outside given rain = 10%
e = total probability that the cat is outside

h|e = (.3 * .1) / (.3 * .1 + .7 * .5) = < 8% chance it is raining, given that the cat is out.

The denominator above represents (the probability of rain * probability of cat out in rain
+ probability of no rain * probability of cat out with no rain)
= total probability of cat being out.

A short hand version of Bayes’s formula that can be derived from it is the “rule of succession”. If we have two possible outcomes, and we start with a 50% chance of each, then as we observe individual successful outcomes, the probability that the next outcome will also be a success increases. The formula here is

n = number of trials
s = number of successes

(s + 1) / (n + 2) = probability of success on the next trial.

If we are reaching in to a dark bag, and our only two possibilities are a red marble, or a “not-red” marble, then by enumeration we start at a 50% chance for each possibility. If we then observe 1 red marble on the first pull, then the chance that the next marble will be red is:

n = 1
s = 1
(s + 1) / (n+2) = 2/3rds

In practical real-world applications of Bayesian analysis, where accurate prior probabilities may be too difficult to obtain, it can be legitimate to introduce a completely subjective guess. If the bias is small and the information set large, this is not a problem. Here, however, with the possibility of strong biases and little data, I will avoid the use of a subjective guess, if at all possible. And, if a guess is needed I will attempt to guess conservatively, provide justification for the guess, and examine its impact on the result. In general, however, the only techniques I will attempt to employ are enumeration of relevant possibilities, the rule of succession, Bayes’s formula, and the standard rules of probability.


Salt, fire and sacrifices

R.T France3 (p. 383) writes, “Apart from general considerations on the metaphorical uses of both fire and salt in biblical literature (each of which yields a variety of possible lines of interpretation, but it is the use of the two together that which is unusual and arresting), the most promising line of approach is via Lv. 2:13, the requirement that grain offerings (which were burned) must be accompanied by salt, together with the more sweeping generalization, ‘With all your offerings you shall offer salt’.” And then he writes, “To be ‘salted with fire’ seems to evoke the imagery of temple sacrifice, but the victims who are ‘salted’ are now the worshipers themselves.”

Leviticus 2:11f (New Jerusalem) reads, “None of the cereal offerings which you offer to Yahweh must be prepared with leaven for you must never include leaven or honey in food burnt for Yahweh. You may offer them to Yahweh as an offering of first-fruits, but they will not make a pleasing smell if they are burned on the altar. You will put salt in every cereal offering that you offer, and you will not fail to put the salt of the covenant of your God on your cereal offering”

What is the probability that Mark wants to call to mind Leviticus and Temple sacrifices here? We’ll start with an enumeration of relevant possibilities. First of all, salt and fire together uniquely identify this piece of Hebrew scripture. We need not consider other passages that the combination could refer to. We’ll consider the possibility that Mark may have wanted to refer to more than one passage in a moment. Secondly, Mark might have no cogent purpose at all in bringing together salt and fire. A third relevant possibility is that salt and fire have been brought together because of their destructive properties, this leads to the cogent idea that sinners will be destroyed by the fire. Another cogent possibility suggested by James R. Edwards2 (pp. 295-296) is the fact that both can purify. This could be related to a purifying torment after the resurrection, or a testing in this life. An imaginative suggestion is that both fire and salt could be used to cauterize the wounds left from the removed body parts. (J.D.M. Derrett, Theology 76 (1973) 364-68) Finally, what if Mark did not want salt and fire to indicate one thing? What if he had completely different purposes for each? One problem with this, is that given the number of possible meanings of each individually, and the number of possible scriptural reference of each individually, anyone listening would have no chance of constructing a meaning. If we were to consider some of these possibilities that seem to yield cogent results, we would have to multiply each possibility by the probability that Mark meant that specific meaning of fire and that specific meaning of salt. For example if fire and salt each had 10 meanings then a combination of any two that were unrelated would have to be multiplied by a factor of 1/100. But while all these possibilities are tiny, collectively they should be one of our enumerated possibilities. 

That gives us six possibilities with a starting 17% probability each (A non-cogent text, Sacrifices/Leviticus, destruction, purification, wound-healing, and obscure separate references). But now we can note that Mark gives us other clues pointing to Leviticus. Mark has given us a reference to Gehenna, a place where human sacrifices by fire were made, (See Jeremiah 7:31 and 2 Chronicles 28:3 for example), giving us a link to sacrifices in Leviticus. Mark gives us images of cutting off body parts, suggesting complete dedication to God, and sacrifices are fully consumed when dedicated to God. Finally we note that salt in Leviticus makes sacrifices acceptable to God, and that Mark gives us text indicating things and people that are acceptable or unacceptable to God. There is punishment for those that cause loss of faith, and reward for those that give a drink of water. We are also told that those that are not against us are for us.

Each of these three things would be unlikely to occur simply by chance here, if Mark was not evoking Temple sacrifices. But these things in combination, salt, fire, the sacrificial fires of Gehenna, dedication to God, and acceptability to God make it an overwhelming probability that Mark wants us to read Leviticus and sacrifices here. We can do the math with a conservative estimate here. What is the chance that a given theme would come up, just by chance, if we were not talking about sacrifices? Based on the frequency of each theme in other similar sized units of biblical text we would have to say the probability is small. We’ll put the chance that one of these themes would occur by accident at less than 5%, and round up to 5% to avoid trying to enumerate the percentage of passages that contain a specific theme. If we also say there is a 50% chance, by enumeration (yes/no), that Mark would choose to use this image if he was talking about sacrifices, we can calculate as follows.

P( Leviticus reference given theme) = P(Leviticus reference apriori) * P(theme given Leviticus reference) / P(theme) =

.17 * .5 / (.17 * .5 + .83 * .05) = 67%

Two more iterations for the other two themes (beyond the salt and fire that uniquely identified the passage) gives us:

.67 * .5 / (.67 * .5 + .33 * .05) = 95%

.95 * .5 / (.95 * .5 + .05 * .05) = 99.5%

This establishes Mark’s use of Leviticus here as a near certainty. But since we did not enumerate passages, what would happen if we picked a higher background frequency for these themes? For example, given that we already have fire, might that give us a fair probability of a Gehenna reference without a sacrifice theme? Of course Hebrew scripture contains over 300 references to fire, and only a handful of references to Gehenna, so this does not seem to be a promising line of argument. But suppose we said there was a 20% chance that each of these themes would occur at random without a sacrifice theme. We would also have to then raise the probability that these would be in Mark if he did mean us to look at Leviticus. A rising tide of background frequency raises all boats. Rather than 50% we would have 1-((1-50%)*(1-20%)) = 60%. If we then repeat our iterative calculation we get an 84% probability of a Leviticus reference here. Even given extremely unfavorable estimates, a Leviticus reference is still quite probable. However going forward, based on my original estimates of less than a 5% chance for each of the themes, I will consider Mark’s reference to Leviticus to be established.

We are not very far out of step with the literature here. Edwards2 (p. 296) has “The above interpretations are generally valid, but they fall short of explaining the presence or meaning of fire and salt at the conclusion of Mark 9. Since v.49 occurs only here and in no other Gospel, it must hold special significance for Mark. The most promising interpretation of vv. 49-50 is to understand them against the background of temple sacrifices, in which both fire and salt played indispensable roles.” France3 (p. 383) points to the ideas of salt and fire combined in order to conclude that Leviticus is the most promising line of approach. By also noting that Gehenna was a place of human sacrifice as a connection, and the connections between the themes of dedication to God, and acceptability to God, and sacrifices, I have argued that Mark is nearly certain to be referring to Leviticus here.

Acceptable or dedicated?

At this point I depart from the literature. The literature focuses on sacrifices as dedicated to God, that is we should make sacrifices, and I focus on sacrifices as acceptable to God, that is we should become metaphorical sacrifices ourselves. To establish initial plausibility of the idea that Mark might have wanted us to read people as metaphorical sacrifices here, we can look to Paul. He says in Romans 12:1 (New Jerusalem). “I urge you, then, brothers, remembering the mercies of God, to offer your bodies as a living sacrifice, dedicated and acceptable to God” [QUSIAN ZWSAN AGIAN EUARESTON TW QEW] (‘Sacrifice, living, holy, well-pleasing to God’).

There is no clear reason to favor one or the other meaning given the text of this pericope prior to v. 49. (Again, please note the pericope begins at v. 38 with John’s question, and then contains an unbroken speech by Jesus). The themes of dedication and acceptability are both there. In favor of my case, I might argue that dedication and sacrifice are more nearly synonymous than sacrifice and acceptability. So “acceptability” in addition to “sacrifice” is a clearer statement of Mark’s intentions beyond sacrifice than “dedication” in addition to sacrifice. On the negative side, one could argue that the dedication theme is a little more clearly presented in the text than the acceptability theme. Both judgments would be subjective, and to avoid that we will give each a 50% probability. Could Mark want us to get both ideas? Yes, of course. But we get different interpretations of the salt metaphor, depending on which theme we choose to use for interpretation. Before trying to set up a probability calculation, however, I’ll just run through the arguments.

First of all, in what way does salt relate to sacrifices in Leviticus? Sacrifices are dedicated to God, but it is the salt that makes them acceptable to God. Therefore the logical relation between metaphorical sacrifices and metaphorical salt is that the salt makes the sacrifices acceptable.

Secondly there is an issue of cogency. “Acceptability” yields a cogent meaning for all occurrences of salt. “Where their worm does not die and the fire is not quenched everyone will indeed be salted (made acceptable to God) by fire. The salt (that makes sacrifices acceptable to God) is good. Now, if salt (what makes a sacrifices acceptable to God) were to become salt-less (unacceptable to God) in nature, in whom will it be seasoned (preserved, renewed)? Have salt (what makes one acceptable to God) in you, and be at peace with one another.” By using “acceptability” for salt, we can see that Mark has written a little riddle. We are left asking “What might God change?” “What might become un-acceptable to him?” “How do we have ‘acceptability to God’ in us now?”

“Dedicated” can yield some cogent results too. “Everyone will become dedicated to God by fire” could refer to trials of discipleship.  “Have a dedication to God within you” works too. But “Now, if dedication to God were to lose its dedication, in/by what/whom will it be seasoned?” does not work. If one negative result served to falsify a hypothesis, I would consider the “dedication” hypothesis to be falsified here. But as Kuhn4 successfully argued, refutation need not always falsify a hypothesis. Additional ad hoc hypotheses can be introduced to explain the anomaly. A different example may be useful here. Suppose we see a sequence – red sphere, red sphere…, and we pick two hypotheses, “red”, and “sphere”. Then we see – “red sphere”, “red cube”, “red sphere”. Is the “sphere” hypothesis wrong? Maybe. But we can claim it is correct, and create a sub-hypothesis to explain the cube anomaly. The “red” hypothesis, however, needs no adjustment. A Bayesian calculation here would show that the “red” hypothesis was more likely to be true than the “sphere” hypothesis. If we had to bet on the next object, we should bet on “red”.

Commentators have added ad hoc hypotheses to explain Mark’s second salt saying on the “dedicated” hypothesis. For example the middle salt saying could just have wound up here because of key word association. Sometimes the middle saying seems to be ignored in the explanation of the passage, after focusing on Leviticus. Moloney7 gives this, “Once this salt, giving a sense and flavor to the Christian’s commitment to the way of Jesus, is lost, nothing can replace it. Whether or not this happens, or how it might happen, is irrelevant; the image retains its power, as one cannot imagine what a salted object might be like without its saltiness”. That works with the general theme of the passage, on the “dedicated” hypothesis, but not with Mark’s wording. Mark says “Now if salt were to become salt-less in nature”, he does not say “Now, if a person were to become salt-less in nature”. So, while “acceptability” gives us cogency for all occurrences of salt, “dedication” only succeeds in 2 of the 3 salt sayings.

Finally we can look at language. Where ‘their worm does not die and the fire is not quenched' everyone will indeed be salted with fire. [OPOU...PAS GAR PURI ALIAQHSETAI]. Strong8 says “GAR” is used both as an intensifier and to assign a reason or explanation. Ignoring this connotation of GAR can allow the worm and the fire to be part of the previous sentence. But connecting the worm and the fire to salting by fire, allows GAR to say that the worm and the fire quoted from Isaiah give a reason for the statement that “everyone will be salted with fire”. If we allow GAR to have this meaning then “dedication” for salt no longer works. This can no longer be about a testing of disciples. Also, as a side note, while Isaiah was referring to death and ultimate indignity with this imagery, Mark may have employed it in an entirely different way than what Isaiah meant. France3 writes (pp381-382) that Is. 66:24 formed the “basis for later Jewish concepts of ‘Gehenna’”, “a term used in apocalyptic literature for the ultimate place of punishment for the ungodly”, (See also Jdt. 16:17 and Ben Sira 7:17).

How should we assess probabilities here? Since we have two opposing hypotheses something that tends to confirm one tends to disconfirm the other. So we’ll award “points” here. “Acceptability” gets a point for logic and a point because “dedicated” does not work with the logical relation of salt to sacrifices in Leviticus. “Acceptable” gets three points for cogency in all three salt sayings and one more point because “dedicated” lacks cogency in one of the salt sayings. On the other side, “dedicated” gets two points for cogency in two of the salt sayings. “Acceptable” get one more points for better preserving the sense of the Greek GAR. We said that each check for logical fit and cogency could count for two “points”, since it is possible that both meanings, “acceptable” and “dedicated” could fit, or not fit, and therefore we had two checks. But for word use it is not possible for both ideas to be better than each other. Therefore word use only counts for one “point” per check. The total score is “acceptability” 7, “dedicated” 2. Then using the rule of succession we get the probability of “acceptable” as the correct meaning here as:

(s+1)/(n+2) = (7+1)/(9+2) = 8/11 = 73%

If this seems low based on the arguments above, then we need to step back and look at what we’ve done here. I made no assumption about how probable it was that Mark would be logical or cogent. But now that we have worked out the probabilities for “acceptable” and “dedicated” we can use these result and see that we would now claim the probability that Mark was logical in his use of metaphor is only 73%. Perhaps we should argue that the whole is greater than the sum of the parts here. If we can read an author to be logical, cogent, and making standard use of language, or read him as not being logical, being less than cogent, and not making use of the standard meaning of language, should we really say there is only a 73% chance that the former reading is correct? This involves incorporating our prior knowledge of how authors generally behave. To use this information in a Bayesian fashion, we would have to enumerate how often Mark (or authors in general perhaps) were logical, cogent, etc. I don’t plan to do that here, but I will make the somewhat subjective claim that I believe it is much higher than 73% of the time.

What if we had said that authors are cogent and/or logical 95% of the time? If we start with probabilities of 1/3 vs. 2/3 based on word usage, as above, and skip consideration of the first and last salt saying where both hypotheses passed, and also suppose that a false hypothesis has a 50/50 chance of success, we get:

P (“acceptable” hypothesis given failure of “dedicated” hypothesis) =

.67 * .5 / ((.67 * .5) + (.33 * .05)) = 95%

Reiterating for the other event gives:

.95 * .5 / ((.95 * .5) + (.05 * .05)) = 99.5%

In this case the “acceptability” hypothesis is firmly demonstrated. But even without assuming that authors are very likely to be cogent and logical, we still have demonstrated that “acceptability” is the most probable reading for “salt”.

We could also continue to marginally increase the probable truth of the hypothesis by looking at Mark 8 where Jesus warns of the yeast of the Pharisees. (Mark 8:14-15) "Be careful," Jesus warned them. "Watch out for the yeast of the Pharisees”. Mark again seems to be using the language of sacrifices from Leviticus as a metaphor for what is inside people. If we were to set up the calculation here, we would argue that if Mark has used the sacrifice metaphor and salt as a positive, there is a 50% chance (by yes/no enumeration) that yeast as a negative would appear elsewhere in Mark. But, given that yeast is in eight books of the Old Testament, the expected “background frequency” of yeast is not that low, and the contrast between what we would expect if our hypothesis was true, and what we would expect if it were false is not extreme, so this would only make a small positive adjustment in our probability.

As a side note, France3 points out (p.382) that "it is a matter of doctrinal debate whether the unquenchable fire refers to conscious torment of those committed to it or to a fire that destroys but never goes out. In support of the latter idea, one could note that Mark is using a metaphor of human sacrifice, and the sacrificial alter fires in the temple were never put out (Lev. 6:9, 13; 9:24)". However, given our argument that the fires makes those committed there acceptable to God, we should also consider the idea that Mark is using Gehenna as a metaphor for a sort of purgatory.

Answering the riddle

Now we need to answer Mark’s riddle. The text of the riddle itself and our knowledge of Christianity in general is enough to suggest the answer involves a new or renewed covenant in Christ. We can then find support for this answer in the larger pericope and in Leviticus. First, in answer to the question “What might become unacceptable?” we have the answer suggested by Leviticus, “the covenant of salt”. Next, we can note the repetition in Mark, beginning at the end of the previous pericope of, “in my name”, “in your name”, “in my name”, “because you are in the name of Christ”, which leads to, “In whom will it be seasoned (preserved, renewed)?”

I think we should consider the “in your name”, “in my name” pair that both refer to demons as one occurrence. But the other two are separate occurrences. There is welcoming little children in his name, and receiving water “because you are in the name of Christ”. Outside of this section, Mark only has one other occurrence of “in my name”, in Mark 13:6. “Many will come in my name saying ‘I am’” (POLLOI ELEUAONTAI EPI TW ONOMATI MOU LEGONTES OTI EGW EIMI), “and they will deceive many”.

For the calculations, first we can look at Leviticus. If we are correct then we are meant to get the “covenant of salt” meaning from Leviticus. If we are wrong at this point, the only other alternative is that we are only talking about sacrificial salt here. Then we would have to read “If sacrificial salt were to become unacceptable in nature, in/by what/whom will it be seasoned (renewed)? This would then be a message about renewing the practice of Temple sacrifice. Given our prior knowledge of Christianity, a message about a renewed covenant is far more likely than one about a renewal of temple sacrifices. This would be difficult to quantify, however.  The best approach might be to say that the rest of Mark contains nothing about renewed temple sacrifice, but there is something in Mark about the covenant. “This is the blood of me, of the covenant” (TOUTO ESTIN TO AIMA MOU THA DIAQHKHS). If we say that given the hypotheses are somewhat contradictory, there is less than a 5% chance that the wrong hypothesis would be confirmed elsewhere in Mark. And if we also say there is a 50% chance that Mark would give us confirmation of the correct hypothesis, then we can calculate –

P = .5 * .5 / (.5 * .5 + .5 * .05) = 91%

Then given that Mark uses “in my name” in only one of 15 other chapters, we can say that based on the background frequency, there is less that a 10% chance that we would expect “in my name” here just by chance. We can also start with a 50% chance that Mark would provide us this clue, if our hypothesis is correct. We need to adjust this upwards for the background frequency. Rather than 50% we have 1-((1-50%)*(1-10%)) = 55%. Then applying Bayes’s formula we get –

P (hypothesis given “in my name”) = .92 * .55 / (.92 * .55 + .08 * .1) = 98.4%

We have to be careful about the calculation for the recurrences of “in my name”, however. Given that “in my name” has already occurred once, we might give it a 50% chance that it would occur again, even if our hypothesis is wrong. If our hypothesis is correct however, after one confirmation, by the rule of succession, we would place the probability of the next confirmation at 2/3rds. Our update formula then is –

P =. 984 * .667 / (.984 * .667 + .016 * .5) = 98.8%

And then reiterating once more for the 3rd occurrence –

P = .988 * .75 / ( .988 * .75 + .012 * .667) = 98.9%

Finally, Mark gives us a disconfirmation for the hypothesis that we might be talking about a renewal of temple sacrifices. Mark 12:33 “To love him with all your heart, with all your mind and strength, and to love your neighbor as yourself, this is far more important than any burnt offering or sacrifice”. Another calculation like the one we used for Mark’s mention of the covenant gives us –

P = .989 * .5 / (.989 * .5 + .011 * .05) = 99.89%

It is highly probable that we are talking about the salt of the covenant and a new or renewed covenant in the name of Christ.

Matthew, Luke, and Salt

Matthew and Luke do not have the central salt saying in the same context, and in their context it clearly means something different. We cannot directly infer from one author’s use of a saying that another author must have meant the same thing. Still, all things being equal, that is what we would tend to expect. Should we adjust our probability down for this? I argue no, because in the way Matthew and Luke revise Mark in other places, we can detect a pattern that might actually lead us to predict that Matthew and Luke would treat Mark’s salt saying in a similar fashion.

There are a number of passages in Mark that could have been used by someone who wanted to argue that the Law of the Old Covenant was no longer valid. Here we are not concerned with what Mark meant to say, but with what some, say late first century proto-Gnostics for example, might have tried to read there.

I do not believe Mark intended to say that the Old Covenant Law no longer applied. A renewed covenant seems more likely. But Mark’s salt sayings could easily be read that way by someone trying to make that case. If we look ahead to Matthew, we find that the salt saying has been completely re-contextualized. Matthew 5:13 "You are the salt of the earth. But if salt were to become tasteless/foolish, in/by what/whom will it become salty? (EAN DE ALAS MWRANQH EN TINI ALISQHSETAI)  It is no longer good for anything, except to be thrown out and trampled by men.” Matthew changes “salt-less” to tasteless/foolish, and changes “seasoned” to “salty”. If we use the “acceptable” reading here we get, “You make the earth acceptable to God. But if you who are the salt were to become foolish in whom will you be made acceptable to God?” Now it is a warning not to stray from the fold. Matthew soon follows this with Matthew 5:17 “Do not think that I have come to abolish the Law or the Prophets; I have not come to abolish them but to fulfill them. I tell you the truth, until heaven and earth disappear, not the smallest letter, not the least stroke of a pen, will by any means disappear from the Law until everything is accomplished.” In Matthew’s context it is very clear that the salt sayings are not about the abolition of the Old Covenant Law.

Also of interest here is Matthew’s choice of the word “MWRANQH” (or "Moranthe"). It has been argued that Matthew got this from an Aramaic saying source which had a word with the root “TPL”, and meant “becomes tasteless”20. The theory continues that Matthew then mistranslates it to mean “becomes insipid”, and thus chooses to use the Greek word “MWRANQH”, “foolish”. However, I would argue that Matthew wrote “foolish”, because he meant “foolish”, that is he was deliberately re-writing Mark. If that was the case, then Matthew would have most likely been well aware of the similarly between “foolish” and “tasteless” in Aramaic, and probably used this fact to help justify his changing the gospel of Mark’s “salt-less” to the gospel of Matthew’s “foolish” and the re-contextualization of the salt saying.

Luke also removes the saying from Mark’s context and makes it a saying about discipleship. “In the same way, any of you who does not give up everything he has cannot be my disciple. Salt then is good. But now if it salt should become tasteless in/by what/whom will it be seasoned? (EAN DE KAI TO ALAS MWRANQH EN TINI ARTUQHSETAI)  It is fit neither for the soil nor for the manure pile; it is thrown out. “

Now we can look for other old/new contrasts. We can look, for example, at the stories at sea. In Mark 4:35-41 the waters obey the command of Jesus. Then in Mark 6:45-52 Jesus comes walking on water through the storm. The disciples may have some trouble seeing him, “he looks like a ghost”. We are then told the disciples have not understood (about the bread), and that their “hearts have been hardened”. In 6:49 “he wanted to pass by them” (PARELQEIN). Then in 6:50 Jesus says “I am”. By the references to a water miracle and “hardening hearts” and other language of Exodus, Mark is evoking Exodus. Now in the two miracles at sea we can see first Moses commanding the waters of the red sea in Exodus, and then the new Exodus of Isaiah 43:16f “Thus says Yahweh who made a way through the sea, a path in the raging waters…Look I am doing something new, it emerges; can you not see it?”

Mark seems to have another reference to Exodus not long before the walking on water. In the sending of the twelve Mark 6:8 says that they should take only sandals a staff, and no spare tunic. This echoes Exodus 12:11 “This is how you are to eat it: with your cloak tucked into your belt, your sandals on your feet and your staff in your hand. Eat it in haste; it is the Lord's Passover.” Mark has sent forth the 12 disciples, in the new Exodus, as the 12 tribes were sent forth in the original Exodus. We can also note that this new Exodus follows shortly after Jesus the prophet is rejected by his own people in Mark 6:1f.

Clearly, if someone wanted to make a case that the Old Covenant had been replaced, they could look here to build that argument. But, Matthew and Luke both remove the Exodus reference; they both say that nothing should be taken on the journey. Matthew has Jesus walking on the water (Mt. 14:22f) but removes Mark’s “hardened hearts”, the “pass them by”, and the exhortation for us to look into this deeply, and see the new Exodus. Luke’s solution is to completely omit the “walking on the water” section.

We can also look again at the “yeast of the Pharisees”. Mark 8:15f – ‘"Be careful," Jesus warned them. "Watch out for the yeast of the Pharisees and of Herod”…” Aware of their discussion, Jesus asked them: "Why are you talking about having no bread? Do you still not see or understand? Are your hearts hardened? Do you have eyes but fail to see, and ears but fail to hear? And don't you remember? When I broke the five loaves for the five thousand, how many basketfuls of pieces did you pick up?" "Twelve," they replied. "And when I broke the seven loaves for the four thousand, how many basketfuls of pieces did you pick up?" They answered, "Seven." He said to them, "Do you still not understand?”’

Here we could have 12 for the 12 tribes of Israel, and 7 for the 7 days of creation. The first feeding is in Jewish territory is for the 12 tribes of Israel, and then the second that takes place in Gentile territory is for all of creation. We could be seen to be moving from the 12 loaves of the bread of the Old Covenant. Leviticus 24:5f "Take fine flour and bake twelve loaves of bread…This bread is to be set out before the LORD regularly, Sabbath after Sabbath, on behalf of the Israelites, as a lasting covenant.” to the bread of the new or renewed covenant, Mark 14:22 “While they were eating, Jesus took bread, gave thanks and broke it, and gave it to his disciples, saying, ‘Take it; this is my body…This is my blood, the blood of the (new) covenant .’” 

We can also note the number of loaves and fish they start with in the miracle of the 5000. There are 5 loaves and 2 fish. In Exodus 16:4 we have “Look, I will rain down bread for you from the heavens (for 5 days)…On the sixth day however…this must be twice as much as they collect on ordinary days.” The numbers 5 and 2 in Mark evoke the 5 days of mana with twice as much on the sixth day from Exodus. So we have another reference to bread in the Old Covenant. 

Given our “salt” reading, Mark first tells us that the Pharisees have lost their “acceptability to God”, they now have “leaven” in them, and then one could try to argue by extension we should understand that bread of the Old Covenant is now filled with leaven, and is unacceptable. Mark may simply have meant a renewed covenant that had included the Jews, but now included everyone, but it would be easy to read an invalidation of the Old Covenant Law here. 

Matthew has Mark’s discussion of the yeast of the Pharisees (Mt. 16:5f), but removes Mark’s references to 12 and 7. Again, Matthew does not want to declare the Old Covenant void. Matthew also changes “Herod” to “the Sadducees” and tells us that the yeast means we should beware of their “teachings”, obscuring Mark’s message that they have become unacceptable to God, and thereby obscuring Mark’s potential rejection of the Old Covenant as unacceptable to God. Matthew even suggests that yeast is a good thing in Matthew 13:33. Again Luke’s solution is to drop one of the feedings entirely.

Based on this pattern of rewrites and omissions in Matthew and Luke, it is not surprising, and in fact we might expect, that Matthew rewrites Mark’s Gehenna section, and re-contextualizes his salt sayings, and that Luke completely omits Mark’s Gehenna section. Their reinterpretation of “salt” should not be used to argue for what Mark meant by “salt” originally.

Connection to the gospel of John

This pericope in Mark is the only one where a question is asked by John in a solo role. The theme of being “in Jesus” is not typical of most of the text of the synoptics; however, it is very characteristic of the gospel of John. (for example, 6:35, 6:36, 7:38, 10:38, 11:25, 11:26, 12:44, 12:46, 14:1, 14:10(x2), 14:11, 14:12, 14:20, 15:2, 15:4, 15:5, 15:6, 15:7, 16:9, 16:33, 17:20, 17:21, 17:23). France3 also points out that “the idea of Jesus as one ‘sent’ is more typical of John than of Mark, and occurs only here in this gospel (except parabolically in 12:6).”

If we assume that this question is not attributed to John by coincidence, and if we agree that the gospel of John is a much later work than the gospel of Mark, then how should we account for this? It seems unlikely that the gospel of John would become attributed to John, based on the association of John with this material in Mark, although we can not count this out completely. However, the arrow of causality would seem to more plausibly point in the other direction. If this is the case, then we would have to say that at whatever point in time this question became associated with the name John in the gospel of Mark, there was already a separate tradition like this associated with the name John. Either there was some earlier version of the gospel of John or its predecessor in circulation, or an oral tradition associated with John was in circulation at the time Mark was written, or this question only came to be associated with John in Mark’s gospel at a latter date, due to redaction. Although this last possibility is rendered more unlikely by the fact that in Luke’s gospel this question is still attributed to John, while the connection to salt has been removed in Luke’s gospel. A redaction scenario would have a difficult time accounting for the text of Luke here. However, any of these scenarios are possible, and in any case, the association seems unlikely to be coincidence.

The salt of Jesus

The salt pericope ends with “
Have (my) salt [hALA] in yourselves, and be at peace with each other “. The first thing to note here is that it is not completely clear that we should read “my” here.  In my text notes, I speculate that hALA might be a “case-less” salt. It could then intentionally be read either as “Have salt in you” (accusative) or “Have my salt, or the salt of one of whom I speak, in you” (genitive). This reading of hALA is made more probable by the preceding question, “In whom will (the covenant) be renewed?”

If we don’t read “my” here, then given the previous arguments the meaning of the end of this pericope is just “have acceptability to God in you”. But if we do read “my” here, then we have one more question to answer, “What can it mean to have the ‘salt of Jesus’ in one?” We could read this as “have in you, that which made Jesus an acceptable sacrifice to God”. Identifying what this something is, however, supposes that we can identify what the author of Mark believed God’s purpose for Jesus was. Some candidates might be “have in you, faith, love, forgiveness, humility, or the Spirit”. But Mark does not tell us that he specifically means any of these.

We might also read “have (my) salt in you”, as meaning “have my acceptability to God in you”.  Or put another way, “have acceptability to God in you, by having me in you”. In support of this reading we can note that we have already argued for a renewed covenant in the name of Jesus, and for reading an instruction here to have “the acceptability to God” of Jesus in oneself. To this we can also add that just before the beginning of the pericope, in Mark 9:37 we have “
…anyone who welcomes (or receives, DECETAI) me, welcomes not me but the one who sent me.” In commenting on Mark 9:37 France3 writes “it is likely that already the concept of receiving God through receiving Jesus would have carried some of the Christological weight it achieves in Jn. 14:6-11, 20-24”.  (John 16:20 reads “On that day you will know that I am in my Father, and you in me, and I in you”) Finally we can note that this pericope, introduced by a question from John, seems to have some sort of an association with similar material in the gospel of John.

So, taken together we have “a renewed covenant in Jesus”, “receive Jesus to receive God”, “have the ‘acceptability to God’ of Jesus in oneself, and an apparent connection to the gospel of John where the instruction to “have Jesus in oneself” is explicit. This, to me, seems to be enough to justify reading the end of the salt pericope as an instruction to “have Jesus (and his acceptability to God) in oneself”. The message of the pericope then is something like “The covenant has been renewed in Jesus. To be acceptable to God, to inherit life in the coming Kingdom of God, and to avoid being made acceptable to God by fire, have Jesus in you, and become metaphorical sacrifices (see also Mark 8:34), dedicated to God”.

A connection to the Eucharist?

We could speculate that the instruction to internalize Jesus might be related to the Eucharist. Making this reading more probable, we can also note that the salt pericope is surrounded by others that could potentially be read as Eucharistic, or could have had a Eucharistic meaning at one time, say in an earlier version of Mark. For example, 9:37 says, “whoever receives me, receives not me, but the one who sent me”. This could clearly be read as Eucharistic.

Then in Mark 10:7-8 we have “
For this reason a man shall leave his father…and the two flesh shall become one”. Removed from its current context, this could be read to mean, “For this reason, a man (Jesus) shall leave his Father (God), and the two flesh shall become one (the Eucharist). And in support of the idea that the gospel of Mark or its ancestor may have at one point advocated “becoming one flesh with Jesus”, we can note that Mark 3:19 identifies Jesus as the “bridegroom”, and Mark 10:7-8 is quoting Genesis, regarding marriage. 

Finally in 10:15 we have, “
whoever does not receive the kingdom of God as a child (of God?), may by no means enter it”. If we put these pieces together, we have something like, “Whoever receives me, receives the one who sent me. Have (my) “acceptability to God” in you. This is why a man shall leave his father…the two flesh shall become one. And whoever does not receive the Kingdom of God, as a child (of God), may be no means enter it.” I think we could speculate here that an earlier version of Mark might have contained an older understanding of the Eucharist, which might have looked like this, “Jesus was a son of God. If you eat his flesh, you will become one flesh with him. Then you will be a son of God too, and can inherit the Father’s Kingdom”. And in support of this idea of followers being children of God, we could point to Mark 3:35 where Jesus identifies those that do the will of God as his true brothers and sisters.

As Jesus became identified as the unique son of God, and as more emphasis was placed on the Spirit, the understanding that involved “becoming one flesh” with Jesus may have been replaced with one involving “becoming one Spirit” with Jesus. Another hint of this early progression of understanding might be in Paul, I Corinthians 6:16-17 “…”
the two, as it is said, become one flesh. But anyone who attaches himself to the Lord is one spirit with him”. By comparing the “becoming one flesh” view with “becoming one flesh with a prostitute”, Paul could be subtly speaking against an older view of “becoming one flesh with Jesus”, in favor of “becoming one Spirit with Jesus”.

Also, in support of a Eucharistic reading, we could point to the fact that some early Christians apparently did practice a Eucharist involving salt.12

There are problems with these arguments however, if we try to read a reference to the Eucharist in the canonical text of Mark. While the surrounding passages here are potentially Eucharistic, or may have been at one point, they either are not Eucharistic in Mark, as it stands, or at least not clearly so. And even if there were early Christians practicing a Eucharist with salt, and even if they were looking here to support that practice, we can not automatically infer that that is what the author of Mark intended. So while I believe the evidence indicates that we can read a potential reference to the Eucharist here, I don’t believe there is enough evidence to render this reading probable in the text of Mark we are presented with.


The purpose here has been to argue for a new reading of Mark’s salt sayings. “Where ‘their worm does not die and the fire in not quenched’ everyone will indeed be salted (purified, made acceptable to God) by fire. The salt (that makes sacrifices acceptable to God) is good. Now, if salt (of the covenant, that makes things acceptable to God) were to become salt-less (unacceptable to God) in nature, in whom will it be seasoned (preserved, renewed)? Have (my) salt (that now makes one acceptable to God) in you, and be at peace with one another.” I believe this reading has been demonstrated to be highly probable.

I have also argued that this can be understood as an instruction to internalize Jesus, in order to internalize his acceptability to God. I argue the message of the pericope then is “The covenant will be renewed in Jesus. To be acceptable to God, to inherit life in the coming Kingdom of God, and to avoid being made acceptable to God by fire, have Jesus in you, and become metaphorical sacrifices dedicated to God”.


1. Danker, William Frederick, et al (BDAG) (2000), “A Greek-English Lexicon of the New Testament and Other Early Christian Literature”, The University of Chicago Press

2. Edwards, James R. (2002), “The Gospel According to Mark” – (“The Pillar New Testament Commentary”), Win. B. Eerdmans Publishing Co.

3. France, R.T. (2002), “The gospel of Mark: a commentary on the Greek text.” – (“The New International Greek Testament Commentary”), Win. B. Eerdmans Publishing Co.

4. Jaynes, E.T. (2003), “Probability Theory: The Logic of Science”, Cambridge University Press

5. Kuhn, Thomas S. (3rd edition 1996), “The structure of Scientific Revolutions”,  The University of Chicago Press

6. Kurlansky, Mark (2002), “Salt, A World History”, Walker Publishing Company Inc.

7. Moloney, Francis J. (2002), “The Gospel of Mark – A commentary”, Hendrickson Publishers, Inc.

8. Strong, James L.L.D.,S.T.D. (1890), “Strong’s Exhaustive Concordance of the Bible”, World Bible Publisher Inc.




12. (14:1) and (4:4)







19. Ptolemy Hephaestion Bk6 (as summarized in Photius, Myriobiblon 190)

20. Black, Matthew (1967), “An Aramaic Approach to the Gospels and Acts”, Oxford University Press

Please send comments to -

Back to Synoptic Main Page